From cd555c9b746a0a21be041fa3938f0ff31d0094fc Mon Sep 17 00:00:00 2001
From: ewowi <ewowi@icloud.com>
Date: Mon, 21 Sep 2026 11:37:57 +0200
Subject: [PATCH 1/5] Add Ethernet board presets and three UI modes
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

A board picks its Ethernet wiring by name instead of eleven pin rows, and the UI now has user, expert and developer modes so a card shows what its reader needs. Adds a FAQ page, and check_docgen becomes a ratchet that refuses a rising warning count per rule as well as on the total.

KPI: 16384lights | Desktop:1924KB | tick:135/89/372/2/6/126/293/22/1454/4/257/68/225/16/21/11/2083/118/34/5/174/47/4us(FPS:7407/11235/2688/500000/166666/7936/3412/45454/687/250000/3891/14705/4444/62500/47619/90909/480/8474/29411/200000/5747/21276/250000) | ESP32:1954KB | src:270(68826) | test:206(45107) | lizard:265w

**Core**
- `NetworkModule` gains an `ethBoard` select over five presets (Classic RMII, Classic RMII (no reset), P4-NANO, S31 CoreBoard, Custom), mirroring `Hub75Driver`'s board-preset shape. A known board writes its map and hides the eleven pin rows; hidden stays bound, so the values still drive the interface. Persisted by label, since the option list is filtered per build and an index would name a different board.
- The preset is applied in `defineControls` rather than `onControlChanged`: the config-restore path fires no change callback, so a saved `ethBoard` would otherwise select a row and never reach the pins.
- Three UI mode levels (`kModeUser`, `kModeExpert`, `kModeDeveloper`) replace the `advanced` flag on `ControlDescriptor`, each level cumulative over the one below.
- `previewsEthernetControls`: the desktop builds the Ethernet controls with no interface behind them, tagged developer-mode, so the presets can be exercised without hardware. Every bring-up path still keys on `hasEthernet`.

**UI**
- The on/off button is built only where `respectsEnabled()` is true; it did nothing on the seven modules that keep ticking regardless.
- The stats line shows from expert mode up, and holding it peeks at the rate instead of latching a mode.
- Distinct glyphs per tier: expert 🎚️, developer 🔧.

**Scripts/MoonDeck**
- `check_docgen` refuses a warning count that rises against the committed report, per rule and on the total. The per-rule half compared internal keys against display labels, so it matched nothing; both key spaces are now the report's labels, and the baseline parser is anchored to the `## By rule` table.
- `check_devices` validates `ethBoard` against the preset table parsed out of `NetworkModule.h`, and binds the three preset rows that restate a chip's `ethConfigDefault` so the two homes cannot drift.
- `check_prose` control fixture said `color` while claiming a British spelling, so it could never produce its two expected alerts and the check failed without ever linting.

**Tests**
- The ratchet test drives `_ratchet` itself; it previously asserted against a local copy of the comparison and passed while the real function compared nothing.

**Docs/CI**
- New FAQ page, every entry linked to the page that answers it.
- A register rule in the documentation standards: no sentence whose job is tone.
- Backlog: the two shipped review findings removed; the QuinLED Dig-Next-2 W5500 wiring recorded, including why RMII is impossible on that package and why SPI boards stay Custom.

**Reviews**
- 👾 Card name, tags and stats sat inside the `respectsEnabled` guard, so every always-on module rendered a nameless card → fixed.
- 👾 The per-rule ratchet compared key spaces that never intersect, shipping a guarantee it did not enforce → fixed, with a test that fails against the old code.
- 👾 The installer read Ethernet as unconfigured on all seven preset boards → `ethConfigured` now accepts a non-Custom preset.
- 👾 The `ethBoard` handler re-ran a rebuild the scheduler had already done → removed.
- 👾 Three preset rows duplicate `ethConfigDefault` with nothing binding them → a drift check, control-checked.
- 👾 Three stale doc lines (a half-finished sentence, `expertMode`, a click gesture) → fixed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 CLAUDE.md                                     |   2 +
 docs/contributing/documentation-standards.md  |   3 +-
 docs/how-to/faq.md                            |  52 ++++++
 docs/moonmodules/core/system.md               |   2 +-
 docs/moonmodules/core/ui.md                   |  10 +-
 docs/reference/metrics/docgen.md              |  13 +-
 docs/reference/metrics/repo-health.json       | 126 +++++++-------
 docs/reference/metrics/repo-health.md         |  70 ++++----
 docs/work/future/backlog-core.md              |  39 ++++-
 mkdocs.yml                                    |   3 +
 moondeck/MoonDeck.md                          |   4 +
 moondeck/check/check_devices.py               |  73 +++++++-
 moondeck/check/check_docgen.py                |  77 ++++++++-
 moondeck/check/check_prose.py                 |   2 +-
 mooninstaller/deviceModels.json               |  52 +-----
 mooninstaller/install.js                      |   8 +-
 src/core/module/Control.h                     |  18 +-
 src/core/module/MoonModule.h                  |   2 +-
 src/core/system/HttpServerModule.cpp          |   5 +-
 src/core/system/NetworkModule.h               | 155 +++++++++++++++--
 src/core/system/SystemModule.h                |   9 +-
 src/platform/desktop/platform_config.h        |   3 +
 src/platform/esp32/platform_config.h          |   3 +
 src/ui/app.js                                 | 157 ++++++++++--------
 src/ui/style.css                              |  20 ++-
 test/python/test_check_docgen.py              |  54 ++++++
 .../scenario_MoonModule_control_change.json   |  26 +--
 .../light/scenario_Audio_mutation.json        |  36 ++--
 test/scenarios/light/scenario_Aurora_fps.json |  50 +++---
 .../light/scenario_Driver_mutation.json       |  24 +--
 .../light/scenario_Effects_composition.json   |   4 +-
 .../light/scenario_Fields_polar_lut.json      |  58 +++----
 .../light/scenario_Fluid_solver.json          |  60 +++----
 .../light/scenario_GridBlacks_blackpixel.json |   4 +-
 .../light/scenario_GridLayout_resize.json     |  12 +-
 .../light/scenario_Layer_base_pipeline.json   |   4 +-
 .../light/scenario_Layer_memory_1to1.json     |   4 +-
 .../light/scenario_Layouts_mutation.json      |  18 +-
 .../scenario_MoonLiveEffect_livescript.json   |  52 +++---
 .../light/scenario_MoonLive_pipeline.json     |  24 +--
 .../scenario_MultiplyModifier_memory_lut.json |   6 +-
 .../scenario_MultiplyModifier_pipeline.json   |   4 +-
 .../light/scenario_Trails_ladder.json         |  34 ++--
 .../light/scenario_modifier_chain.json        |  18 +-
 .../light/scenario_modifier_swap.json         |  12 +-
 test/scenarios/light/scenario_perf_full.json  |  66 ++++----
 test/scenarios/light/scenario_perf_light.json |  12 +-
 .../light/scenario_peripheral_grid_sweep.json |  72 ++++----
 .../light/scenario_peripheral_switch.json     |  16 +-
 test/unit/light/unit_Drivers_livePalettes.cpp |   8 +-
 50 files changed, 1011 insertions(+), 575 deletions(-)
 create mode 100644 docs/how-to/faq.md

diff --git a/CLAUDE.md b/CLAUDE.md
index 6171171f..2779b40c 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -208,6 +208,8 @@ Each name is a script under `moondeck/`, run through `uv run`; the command and w
 **`test_host --ui` is never a gate.** The UI scenario runs drive a real browser against a running device and are what the documentation clips are recorded from, so they cost minutes and skip wholesale without a desktop and a Playwright browser. They run on request only, never as part of pre-commit, pre-merge or pre-release, and the `src/ui` trigger above means `--js` alone.
 
 
+**`check_docgen` is a ratchet.** Errors are resolved before a commit, and warnings may only fall: the committed `docs/reference/metrics/docgen.md` is the number to beat, on the total and on every rule. It fails a run that raises either. Per rule as well as per total, because a total hides one rule paying for another, and because the cheapest way to satisfy a width rule is to split a line, which raises the block count and fixes nothing. A rule whose own limit changed is the one case to say so in the commit.
+
 Three checks earn their place for a reason worth knowing. **Repo health** is the only place the creeping numbers are visible: flash and DRAM per target, binary size, the tick matrix, line counts, complexity warnings. Its diff belongs in the commit and its deltas in the commit message. It runs when the code changes rather than on every commit, because its timings drift with the host: on a docs-only diff it records a regression that nothing in the diff caused. **The no-backend build** catches a helper left unused outside its guard, fatal under GCC while clang stays silent. **ESP32 firmware fresh** compares the binary against every source in a tenth of a second and catches the edit that was never compiled; compile for real after an sdkconfig or toolchain change. The [provisioning path](moondeck/MoonDeck.md#improv_smoke_test) is the five files MoonDeck names.
 
 
diff --git a/docs/contributing/documentation-standards.md b/docs/contributing/documentation-standards.md
index 60becd2d..71eda51d 100644
--- a/docs/contributing/documentation-standards.md
+++ b/docs/contributing/documentation-standards.md
@@ -140,6 +140,7 @@ Two scales below a page. **A module** has exactly one reference page written and
 - **A page is read start to finish by one reader**, a **user** or a **developer**. A user brings no coding and no hardware knowledge beyond plugging in a board; a developer brings C++, embedded and this codebase's shape. Where a page serves both, lead with the user and put the depth lower down.
 - **The headings are the page's table of contents, and they read top to bottom.** A few lines say what the page is and one paragraph says how it is laid out. The sections then follow in the order a reader needs them. A title that makes sense only after reading the body is the order being wrong.
 - **One tone of voice, everywhere: factual, no nonsense.** State what is true and what to do, addressing the reader as "you". Leave out enthusiasm, apology, and how we felt building it. Only the assumed knowledge changes between pages, never the voice.
+- **No sentence whose job is tone.** Every sentence carries a fact the reader needs. Three shapes to cut: a second person used for effect rather than instruction, a flourish before any information arrives, and a rhetorical question the page then answers. Vale checks spelling, sentence length and weasel words, so this one is the writer's judgement and the reviewer's check.
 - **Follow the [principles](../CLAUDE.md#principles).** Three bear on documentation directly:
     - **Minimalism**: every fact has one home; history lives in git.
         - **Present tense only.** "No X anymore" narrates a removal, which is history. Describe the path that exists today.
@@ -258,7 +259,7 @@ The first five cut and the last adds, deliberately: the result is a short line o
 
 **Depth is homed rather than forbidden.** One line is room to say why. Past that the reasoning belongs in an `@moreinfo` appendix, with an `@xref` back from the line that raised the question. Both kinds of file have one. A header's sits on its class or file lead, and an implementation file's on its own file lead, which 212 headers already carry. A larger cap for implementation files was tried and removed, because it homed depth inline. The reasoning then sits beside one call rather than where a reader goes looking for it.
 
-**A finding in a header is an error; one in an implementation file is a warning.** This follows the same line. A header's comments are the published page, so a defect there ships, while a `.cpp` publishes nothing and its comments are a note to the next reader. Both are counted and both are reported, because a warning nobody sees is a warning nobody fixes. The split stages the sweep rather than ranking the two kinds of comment, so it goes and everything blocks once the warning column reaches zero. That is how Vale's own config promotes a page to error as the sweep finishes it.
+**A finding in a header is an error; one in an implementation file is a warning, with one rule in each direction.** No-hard-wrap blocks in both kinds of file: the tree is at zero findings, so there is nothing left to stage. The line-length cap warns in both, because it is new and its findings are lines nobody wrote wrongly. A staged rule joins the others as the tree meets it. A header's comments are the published page, so a defect there ships, while a `.cpp` publishes nothing and its comments are a note to the next reader. Both are counted and both are reported, because a warning nobody sees is a warning nobody fixes. The split stages the sweep rather than ranking the two kinds of comment, so it goes and everything blocks once the warning column reaches zero. That is how Vale's own config promotes a page to error as the sweep finishes it.
 
 Enforced by [`check_docgen.py`](../moondeck/check/check_docgen.py) over every header under `src/`, the vendored ones excepted.
 
diff --git a/docs/how-to/faq.md b/docs/how-to/faq.md
new file mode 100644
index 00000000..38d861ca
--- /dev/null
+++ b/docs/how-to/faq.md
@@ -0,0 +1,52 @@
+# FAQ
+
+Questions people ask on a first run, in the order they meet them: installing, then first light, then the interface they are looking at, then effects, then the driver that reaches real hardware.
+
+A symptom rather than a question belongs in [Troubleshooting](troubleshooting.md), which starts from what you can see.
+
+## Installing
+
+- **Which firmware does my board take?** The web installer picks it from the board you choose. See [Install and first light](../gettingstarted.md).
+- **Can I run it without a board?** Yes, on macOS, Linux and Windows. See [Installing to desktop](installing-to-desktop.md).
+- **Which hardware do the heavy effects want?** The ones computing a field per light cost CPU, so a desktop or a small always-on machine runs them where a microcontroller cannot. See [Installing on Linux](installing-on-linux.md).
+- **How do I update?** Over the air from the Firmware card, or by URL. See [Updating firmware](updating-firmware.md).
+
+## First light
+
+- **Where do I set the number of lights?** On a Layout, not on the driver. See [Layouts](../moonmodules/light/layouts.md).
+- **Where do I add an effect?** Expand a Layer under Effects, then **add module**: a layer is what composites effects and holds the blend mode. See [Build your first light show](../tutorials/first-light-show.md).
+- **Nothing lights up.** See [Troubleshooting](troubleshooting.md#the-lights-are-dark).
+
+## The interface
+
+- **What is the number on a module card?** That module's own tick time, shown from expert mode up. Hold it to peek at the rate, which inverts it: a module ticking in 22 µs reads as 45K fps. The device's frame rate is `fps` on the [System module](../moonmodules/core/system.md#system).
+- **The card is busy, and I do not need half of it.** Set `mode` on the [System module](../moonmodules/core/system.md#system) to `user`, `expert` or `developer`: each control names the mode it needs.
+- **The preview changed resolution.** It follows the layout, so a geometry change moves it. See [Preview](../moonmodules/light/drivers.md#preview).
+
+## Effects and scripting
+
+- **Can I write my own effect?** Yes, in MoonLive, on the device, live. See [Write your first script](../tutorials/first-script.md).
+- **How do I save a look?** Presets. See [Presets](presets.md).
+- **Can I drive it from a control surface or a DAW?** Over MIDI, OSC and DMX. See [Control surface](control-surface.md).
+
+## Drivers and panels
+
+- **Which driver do I need?** One per output kind: LEDs on pins, a panel card over Ethernet, HUB75 direct, or a network protocol. See [Drivers](../moonmodules/light/drivers.md).
+- **Which board drives a panel card?** One with gigabit Ethernet, so an S31 or a desktop. An S3 or P4 is 100 Mbit and wants a gigabit switch between it and the card. See [Panel cards](panel-cards.md#the-one-hardware-fact-that-decides-everything).
+- **My ColorLight card does almost nothing.** These cards need a 1000 Mbps link, and the driver reports the negotiated speed in its status line. See [Panel cards](panel-cards.md).
+- **My HUB75 panel is one column out.** Set `clockEdge` to `falling`: some panel chips sample the shift clock on the other edge. See [Drivers](../moonmodules/light/drivers.md).
+- **The HLS stream lags by seconds.** HLS buffers whole segments, so the delay is the format. For a live view use [Preview](../moonmodules/light/drivers.md#preview) instead of [HLS](../moonmodules/light/drivers.md#hls).
+
+## Network
+
+- **Ethernet and WiFi fight.** Pick one in the [Network module](../moonmodules/core/system.md#network)'s `mode`, because both leaves two default routes.
+
+## Developing
+
+- **What is MoonDeck, and do I need it?** A browser console that builds, flashes, runs, tests and checks the project, and drives the boards on your bench. Everything it offers is a script under `moondeck/`, so the CLI and the console run the same code and neither is required. See [MoonDeck](../explanation/architecture/moondeck.md).
+- **Why not PlatformIO or pioarduino?** The ESP32 build is ESP-IDF-native, tracking IDF pre-releases against a pinned commit for chips like the P4 and S31, and the hot-path drivers use the vendor APIs directly. The tooling also covers far more than compile-upload-monitor. See [Building](building.md#moondeck-the-dev-console).
+- **How do I start?** `uv run moondeck/moondeck.py`, then open <http://localhost:8420>. Build and run the desktop first: it needs no board and proves the toolchain. See [Building](building.md).
+
+## Still stuck
+
+Name the geometry, the driver, the peripheral and the measured refresh. Those four decide which path you are on. See [Log an issue](logging-an-issue.md).
diff --git a/docs/moonmodules/core/system.md b/docs/moonmodules/core/system.md
index 3dde5c47..c5eb0034 100644
--- a/docs/moonmodules/core/system.md
+++ b/docs/moonmodules/core/system.md
@@ -14,7 +14,7 @@ The device's identity and vitals — name (behind mDNS `<name>.local`, the SoftA
 
 - `deviceName` — the identity behind mDNS, the SoftAP SSID and the DHCP hostname.
 - `deviceModel` — the board model (drives the installer catalog entry).
-- `expertMode` — reveals advanced controls (marked 🔧) across the UI; off by default.
+- `mode` — how much of the UI is shown: `user`, `expert` (🎚️) or `developer` (🔧), cumulative.
 - `logLevel` — serial verbosity, defaulting to Warn. The first 60 s always logs at Info.
 - read-only vitals — `uptime`, `fps`, `heap`, `psram`, `flash`, `chip`, and per-module footprint.
 
diff --git a/docs/moonmodules/core/ui.md b/docs/moonmodules/core/ui.md
index 9c070cb8..62c0fbf5 100644
--- a/docs/moonmodules/core/ui.md
+++ b/docs/moonmodules/core/ui.md
@@ -122,11 +122,12 @@ indentation. Nesting depth shows as progressively lighter backgrounds and a left
 - **Help link (?)** at the far right of the title row opens the module's doc page in a new tab. The
   path comes from `docPath` in `/api/types` (engine-provided, relative to `docs/moonmodules/`); omitted
   when the type declares none.
-- **Stats line** — `🕒 <timing>` then `🧠 <static>[ + <dynamic>]`. Timing is fps or µs/ms per the
-  global toggle (µs under 1 ms, ms above), omitted when the module has no measured loop time. Memory is
+- **Stats line** — `🕒 <timing>` then `🧠 <static>[ + <dynamic>]`. Timing is a tick time
+  (µs under 1 ms, ms above), omitted when the module has no measured loop time. Memory is
   the C++ object size (`classSize`); the `+ <dynamic>` part (`dynamicBytes`, heap) shows only when the
-  module allocated heap. Clicking cycles the timing figure fps ↔ ms (persists in
-  `localStorage['mm_timing_mode']`, applies to all cards).
+  module allocated heap. The line shows from expert mode up, and holding it peeks at the rate on every
+  card, the password field's gesture: a tick time inverts to a figure that reads as a frame rate and is
+  not one, so the rate is a press away rather than a mode the cards sit in.
 - **Reorder is drag-and-drop** (HTML5 DnD) on the whole card, desktop and mobile. A drag starting on
   an interactive control is canceled in `dragstart` so the control's own gesture wins; a drag is
   accepted only when source and target share the same `.card-children` container (true siblings).
@@ -265,7 +266,6 @@ The endpoints, the `/ws` frame shape, and the streaming state sink are owned by
 ```text
 mm_selectedRoot     id of the currently-selected root module    (string)
 mm_theme            "dark" | "light"                            (default: "dark")
-mm_timing_mode      "fps" | "ms"                                (default: "fps")
 ```
 
 No other client state persists; reorder, control values, etc. all live on the device.
diff --git a/docs/reference/metrics/docgen.md b/docs/reference/metrics/docgen.md
index 98d61fb4..4b638757 100644
--- a/docs/reference/metrics/docgen.md
+++ b/docs/reference/metrics/docgen.md
@@ -4,7 +4,7 @@ Generated by [`moondeck/check/check_docgen.py`](../../../moondeck/check/check_do
 
 Every place the generated documentation breaks the shape [the standards](../../contributing/documentation-standards.md#the-card) define. Current state only: the trend is this file's git history. The list only shrinks.
 
-**0 error(s)** and **3141 warning(s)** across 217 page(s).
+**0 error(s)** and **3135 warning(s)** across 217 page(s).
 
 An error is in a file that generates a documentation page, a header or a catalog page, so the finding is a defect in what gets published and it fails the gate. A warning is in an implementation file, which publishes nothing: its comments are a note to the next reader, worth fixing without being worth stopping a commit for. Both are counted here, because a warning nobody sees is a warning nobody fixes.
 
@@ -15,11 +15,10 @@ The split is temporary. It stages the sweep rather than ranking the two kinds of
 | Rule | Errors | Warnings |
 |---|---:|---:|
 | over-wide comment lines | 0 | 2229 |
-| multi-line comment blocks | 0 | 606 |
+| multi-line comment blocks | 0 | 604 |
 | over-long sentences | 0 | 108 |
 | over-wide doc lines | 0 | 102 |
 | member deep dives | 0 | 92 |
-| hard wraps | 0 | 4 |
 
 ## By documentation area
 
@@ -27,7 +26,7 @@ The unit a sweep runs in: one summary page and the headers it owns, as the [hier
 
 | Summary page | Errors | Warnings |
 |---|---:|---:|
-| `(tests, no card)` | 0 | 2213 |
+| `(tests, no card)` | 0 | 2207 |
 | `platform/index.md` | 0 | 518 |
 | `core/system.md` | 0 | 410 |
 
@@ -37,7 +36,7 @@ Per area, since that is the unit a sweep runs in: the files ranked, then the rul
 
 ### (tests, no card)
 
-**0 error(s)** and **2213 warning(s)** across 184 file(s).
+**0 error(s)** and **2207 warning(s)** across 184 file(s).
 
 | Findings | File | |
 |---:|---|---|
@@ -51,9 +50,9 @@ Per area, since that is the unit a sweep runs in: the files ranked, then the rul
 | 44 | `test/unit/core/unit_moonlive_codegen_xtensa.cpp` | warning |
 | 42 | `test/unit/light/unit_ParallelSlots.cpp` | warning |
 | 41 | `test/scenario_runner.cpp` | warning |
-| 1-41 each | *174 more warning files, 1547 findings* | |
+| 1-41 each | *174 more warning files, 1541 findings* | |
 
-By rule, warnings: 1672 over-wide comment lines, 241 multi-line comment blocks, 106 over-long sentences, 101 over-wide doc lines, 89 member deep dives, 4 hard wraps.
+By rule, warnings: 1672 over-wide comment lines, 239 multi-line comment blocks, 106 over-long sentences, 101 over-wide doc lines, 89 member deep dives.
 
 ### platform/index.md
 
diff --git a/docs/reference/metrics/repo-health.json b/docs/reference/metrics/repo-health.json
index 6df0cd07..b47355f0 100644
--- a/docs/reference/metrics/repo-health.json
+++ b/docs/reference/metrics/repo-health.json
@@ -1,8 +1,8 @@
 {
-  "commit": "46178a32",
+  "commit": "7bfbfa78",
   "flash": {
-    "esp32s3-n16r8": 2131536,
-    "desktop": 1953560,
+    "esp32s3-n16r8": 2134032,
+    "desktop": 1970568,
     "esp32": 2080464,
     "esp32p4rev1-eth": 2029776,
     "esp32p4rev1-eth-wifi": 2284640,
@@ -21,8 +21,8 @@
     "esp32s31": "2026-09-06",
     "esp32": "2026-09-20",
     "esp32-pico": "2026-09-09",
-    "esp32s3-n16r8": "2026-09-18",
-    "desktop": "2026-09-20",
+    "esp32s3-n16r8": "2026-09-21",
+    "desktop": "2026-09-21",
     "esp32s3-n8r8": "2026-09-08",
     "esp32s3-zero": "2026-09-08",
     "esp32-16mb": "2026-09-09",
@@ -31,20 +31,20 @@
   },
   "perf": {
     "desktop": {
-      "tick_us": 120,
-      "fps": 8333,
+      "tick_us": 135,
+      "fps": 7407,
       "scenario_p50": {
         "Layer_base_pipeline": {
           "p50": 70,
           "p95": 74,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "Layer_memory_1to1": {
           "p50": 5,
           "p95": 24,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         }
       }
     },
@@ -58,7 +58,7 @@
           "p50": 127,
           "p95": 139,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "p50": 89895,
@@ -172,9 +172,9 @@
       "Audio_mutation": {
         "desktop-macos": {
           "p50": 22,
-          "p95": 47,
+          "p95": 38,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 40,
@@ -197,18 +197,18 @@
       },
       "Aurora_fps": {
         "desktop-macos": {
-          "p50": 1526,
+          "p50": 1521,
           "p95": 1947,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         }
       },
       "Driver_mutation": {
         "desktop-macos": {
           "p50": 20,
-          "p95": 23,
+          "p95": 30,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 42,
@@ -234,7 +234,7 @@
           "p50": 144,
           "p95": 160,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 549,
@@ -245,10 +245,10 @@
       },
       "Fields_polar_lut": {
         "desktop-macos": {
-          "p50": 1275,
-          "p95": 2014,
+          "p50": 1263,
+          "p95": 1624,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         }
       },
       "Fluid_solver": {
@@ -256,7 +256,7 @@
           "p50": 217,
           "p95": 250,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         }
       },
       "GridBlacks_blackpixel": {
@@ -264,7 +264,7 @@
           "p50": 2,
           "p95": 2,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "p50": 267,
@@ -290,7 +290,7 @@
           "p50": 120,
           "p95": 142,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "p50": 82231,
@@ -334,7 +334,7 @@
           "p50": 70,
           "p95": 74,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 118,
@@ -348,7 +348,7 @@
           "p50": 5,
           "p95": 24,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 1,
@@ -362,7 +362,7 @@
           "p50": 93,
           "p95": 100,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 111,
@@ -414,7 +414,7 @@
           "p50": 5,
           "p95": 6,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "p50": 8255,
@@ -464,7 +464,7 @@
           "p50": 5,
           "p95": 5,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 1,
@@ -476,9 +476,9 @@
       "MultiplyModifier_memory_lut": {
         "desktop-macos": {
           "p50": 3,
-          "p95": 4,
+          "p95": 3,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 3,
@@ -492,7 +492,7 @@
           "p50": 119,
           "p95": 142,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 225,
@@ -503,18 +503,18 @@
       },
       "Trails_ladder": {
         "desktop-macos": {
-          "p50": 359,
+          "p50": 358,
           "p95": 658,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         }
       },
       "modifier_chain": {
         "desktop-macos": {
-          "p50": 44,
+          "p50": 43,
           "p95": 46,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 69,
@@ -534,7 +534,7 @@
           "p50": 23,
           "p95": 25,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32-eth": {
           "p50": 1010,
@@ -572,7 +572,7 @@
           "p50": 251,
           "p95": 295,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "p50": 16915,
@@ -604,7 +604,7 @@
           "p50": 15,
           "p95": 21,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "p50": 2485,
@@ -646,9 +646,9 @@
         },
         "desktop-macos": {
           "p50": 254,
-          "p95": 309,
+          "p95": 324,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "desktop-windows": {
           "p50": 649,
@@ -674,7 +674,7 @@
           "p50": 4,
           "p95": 4,
           "n": 32,
-          "last": "2026-09-20"
+          "last": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "p50": 217,
@@ -698,54 +698,54 @@
     }
   },
   "loc": {
-    "core": 21849,
-    "light": 29887,
-    "platform": 16325,
-    "ui": 11196,
-    "test": 56157,
-    "moondeck": 26920
+    "core": 21994,
+    "light": 29903,
+    "platform": 16346,
+    "ui": 11223,
+    "test": 56265,
+    "moondeck": 27113
   },
   "comments": {
     "core": {
-      "lines": 5548,
+      "lines": 5575,
       "ratio": 0.28
     },
     "light": {
-      "lines": 7250,
-      "ratio": 0.272
+      "lines": 7262,
+      "ratio": 0.273
     },
     "platform": {
-      "lines": 3470,
+      "lines": 3481,
       "ratio": 0.238
     },
     "ui": {
-      "lines": 3333,
-      "ratio": 0.314
+      "lines": 3350,
+      "ratio": 0.315
     },
     "test": {
-      "lines": 6339,
+      "lines": 6345,
       "ratio": 0.131
     },
     "moondeck": {
-      "lines": 4495,
-      "ratio": 0.19
+      "lines": 4531,
+      "ratio": 0.191
     }
   },
   "tests": {
-    "cases": 2065,
+    "cases": 2067,
     "scenarios": 27
   },
   "docs": {
-    "md_files": 137,
-    "md_lines": 28893,
+    "md_files": 138,
+    "md_lines": 28986,
     "plans_files": 31,
-    "backlog_lines": 2986,
+    "backlog_lines": 3025,
     "lessons_lines": 518,
-    "claude_md_lines": 277
+    "claude_md_lines": 279
   },
   "complexity": {
-    "functions": 3621,
-    "over_threshold": 264,
+    "functions": 3627,
+    "over_threshold": 265,
     "worst_ccn": 128
   }
 }
diff --git a/docs/reference/metrics/repo-health.md b/docs/reference/metrics/repo-health.md
index 54d6d801..851eb4f3 100644
--- a/docs/reference/metrics/repo-health.md
+++ b/docs/reference/metrics/repo-health.md
@@ -1,6 +1,6 @@
 # Repo health
 
-Measured at `46178a32`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
+Measured at `7bfbfa78`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
 
 Current state only; the trend is this file's git history (`git log -p docs/reference/metrics/repo-health.md`). Nothing here fails a build: the numbers make growth visible, the judgment stays human.
 
@@ -8,19 +8,19 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Flash | Capacity | Used | Built |
 |---|---:|---:|---:|:--:|
-| desktop | 1,908 KB | - | - | yes |
-| esp32 | 2,032 KB | 2,496 KB | 81% | yes |
-| esp32-16mb | 2,012 KB | 4,096 KB | 49% | **STALE 11d** |
-| esp32-eth | 1,642 KB | 2,496 KB | 66% | **STALE 9d** |
-| esp32-pico | 2,058 KB | 3,072 KB | 67% | **STALE 11d** |
+| desktop | 1,924 KB (+17 KB) ⚠ | - | - | yes |
+| esp32 | 2,032 KB | 2,496 KB | 81% | carried 1d |
+| esp32-16mb | 2,012 KB | 4,096 KB | 49% | **STALE 12d** |
+| esp32-eth | 1,642 KB | 2,496 KB | 66% | **STALE 10d** |
+| esp32-pico | 2,058 KB | 3,072 KB | 67% | **STALE 12d** |
 | esp32-wrover | 1,801 KB | - | - | carried (age?) |
-| esp32p4rev1-eth | 1,982 KB | 4,096 KB | 48% | carried 2d |
-| esp32p4rev1-eth-wifi | 2,231 KB | 4,096 KB | 54% | **STALE 12d** |
+| esp32p4rev1-eth | 1,982 KB | 4,096 KB | 48% | carried 3d |
+| esp32p4rev1-eth-wifi | 2,231 KB | 4,096 KB | 54% | **STALE 13d** |
 | esp32p4rev3-eth | 1,605 KB | - | - | carried (age?) |
-| esp32s3-n16r8 | 2,082 KB | 4,096 KB | 51% | carried 2d |
-| esp32s3-n8r8 | 2,038 KB | 3,072 KB | 66% | **STALE 12d** |
-| esp32s3-zero | 1,977 KB | 2,496 KB | 79% | **STALE 12d** |
-| esp32s31 | 2,294 KB | 4,096 KB | 56% | **STALE 14d** |
+| esp32s3-n16r8 | 2,084 KB (+2 KB) ⚠ | 4,096 KB | 51% | yes |
+| esp32s3-n8r8 | 2,038 KB | 3,072 KB | 66% | **STALE 13d** |
+| esp32s3-zero | 1,977 KB | 2,496 KB | 79% | **STALE 13d** |
+| esp32s31 | 2,294 KB | 4,096 KB | 56% | **STALE 15d** |
 | qemu | 1,351 KB | - | - | carried (age?) |
 
 `Built: yes` was measured this run. `carried (age?)` was not rebuilt either and predates this record, so its age is unknown: it dates itself on the next build. `carried Nd` was NOT rebuilt and its number is N days old, so an absent delta says nothing about the change. **STALE** marks a carry older than 7 days: the number has gone unchecked long enough that growth will surface later as one jump, blamed on whichever commit happens to rebuild that target. `Used` is against the app slot in the firmware's own partition table.
@@ -29,7 +29,7 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Tick | FPS |
 |---|---:|---:|
-| desktop | 120 µs (−8 µs) ✓ | 8,333 (+521) ✓ |
+| desktop | 135 µs (+15 µs) ⚠ | 7,407 (−926) ⚠ |
 | esp32 | 8,354 µs | 119 |
 
 ### Scenario tick by target (p50 of each sample window)
@@ -37,30 +37,30 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 | Scenario | desktop-macos | desktop-windows | esp32 | esp32s3-n16r8 | esp32p4rev1-eth | esp32s31 | esp32-eth | esp32-eth-wifi | unknown |
 |---|---|---|---|---|---|---|---|---|---|
 | Audio_mutation | 22 | 40 ? | 13,152 | 47 ? | - | - | - | - | - |
-| Aurora_fps | 1,526 (−9) ✓ | - | - | - | - | - | - | - | - |
+| Aurora_fps | 1,521 (−5) ✓ | - | - | - | - | - | - | - | - |
 | Driver_mutation | 20 | 42 ? | 12,812 | 39 ? | - | - | - | - | - |
-| Effects_composition | 144 (−1) ✓ | 549 ? | - | - | - | - | - | - | - |
-| Fields_polar_lut | 1,275 (−7) ✓ | - | - | - | - | - | - | - | - |
-| Fluid_solver | 217 (−1) ✓ | - | - | - | - | - | - | - | - |
+| Effects_composition | 144 | 549 ? | - | - | - | - | - | - | - |
+| Fields_polar_lut | 1,263 (−12) ✓ | - | - | - | - | - | - | - | - |
+| Fluid_solver | 217 | - | - | - | - | - | - | - | - |
 | GridBlacks_blackpixel | 2 | 8 ? | 269 ? | 267 ? | - | - | - | - | - |
-| GridLayout_resize | 120 (−1) ✓ | 219 ? | 1,352 ? | 1,011 ? | 1,143 ? | - | 95,771 ? | 82,231 ? | - |
+| GridLayout_resize | 120 | 219 ? | 1,352 ? | 1,011 ? | 1,143 ? | - | 95,771 ? | 82,231 ? | - |
 | Layer_base_pipeline | 70 | 118 ? | - | - | - | - | - | - | - |
 | Layer_memory_1to1 | 5 | 1 ? | - | - | - | - | - | - | - |
 | Layouts_mutation | 93 | 111 ? | 13,692 | 45 ? | - | - | 27 ? | - | - |
 | MoonLiveEffect_controls | 11 ? | - | 12,901 | 4,624 ? | - | - | - | - | - |
 | MoonLiveEffect_livescript | 5 | - | 13,433 ? | 8,255 ? | 11,336 ? | - | - | - | - |
 | MoonLive_pipeline | 5 | 1 ? | 9,604 ? | 3,278 ? | - | 11,398 ? | - | - | 4,393 ? |
-| MoonModule_control_change | 127 (−6) ✓ | 262 ? | 212 ? | 166 ? | 165 ? | - | 111,731 ? | 89,895 ? | - |
+| MoonModule_control_change | 127 | 262 ? | 212 ? | 166 ? | 165 ? | - | 111,731 ? | 89,895 ? | - |
 | MqttModule_haDiscovery_toggle | 3 ? | - | 36 ? | 36 ? | - | - | - | - | - |
 | MultiplyModifier_memory_lut | 3 | 3 ? | - | - | - | - | - | - | - |
 | MultiplyModifier_pipeline | 119 | 225 ? | - | - | - | - | - | - | - |
 | NetworkModule_eth_reconfigure | - | - | 1,169 ? | 97,843 ? | - | - | - | - | - |
 | NetworkModule_mdns_toggle | 13 ? | - | 36 ? | 36 ? | 21 ? | - | 109,767 ? | 93,963 ? | - |
-| Trails_ladder | 359 | - | - | - | - | - | - | - | - |
-| modifier_chain | 44 | 69 ? | 13,337 | - | - | - | - | - | - |
+| Trails_ladder | 358 (−1) ✓ | - | - | - | - | - | - | - | - |
+| modifier_chain | 43 (−1) ✓ | 69 ? | 13,337 | - | - | - | - | - | - |
 | modifier_swap | 23 | 41 ? | 12,250 | 354 ? | 362 ? | - | 1,010 ? | - | - |
 | perf_full | 251 | 592 ? | 10,392 | 16,915 ? | 17,433 ? | - | - | - | - |
-| perf_light | 15 (−1) ✓ | 35 ? | 2,183 | 2,485 ? | 2,038 ? | - | - | - | - |
+| perf_light | 15 | 35 ? | 2,183 | 2,485 ? | 2,038 ? | - | - | - | - |
 | peripheral_grid_sweep | 254 | 649 ? | 6,991 ? | - | 11,495 ? | 12,273 ? | - | - | - |
 | peripheral_switch | 4 | 9 ? | 437 | 46 ? | 217 ? | - | - | - | - |
 
@@ -81,36 +81,36 @@ These build a bare pipeline with no optional modules, so a change here is a chan
 
 | Area | Lines | Comments | Comment share |
 |---|---:|---:|---:|
-| core | 21,849 (−648) ✓ | 5,548 | 28.0 % (−2.3 %) ✓ |
-| light | 29,887 | 7,250 | 27.2 % |
-| platform | 16,325 (−984) ✓ | 3,470 | 23.8 % (−4.8 %) ✓ |
-| ui | 11,196 | 3,333 | 31.4 % |
-| test | 56,157 (−4,849) ✓ | 6,339 | 13.1 % (−8.1 %) ✓ |
-| moondeck | 26,920 (+106) ⚠ | 4,495 | 19.0 % (+0.1 %) ⚠ |
+| core | 21,994 (+145) ⚠ | 5,575 | 28.0 % |
+| light | 29,903 (+16) ⚠ | 7,262 | 27.3 % (+0.1 %) ⚠ |
+| platform | 16,346 (+21) ⚠ | 3,481 | 23.8 % |
+| ui | 11,223 (+27) ⚠ | 3,350 | 31.5 % (+0.1 %) ⚠ |
+| test | 56,265 (+108) ⚠ | 6,345 | 13.1 % |
+| moondeck | 27,113 (+193) ⚠ | 4,531 | 19.1 % (+0.1 %) ⚠ |
 
 ## Tests
 
 | Kind | Count |
 |---|---:|
-| unit cases | 2,065 |
+| unit cases | 2,067 (+2) ✓ |
 | scenarios | 27 |
 
 ## Complexity
 
 | Metric | Value |
 |---|---:|
-| functions | 3,621 |
-| over threshold | 264 |
+| functions | 3,627 (+6) ✓ |
+| over threshold | 265 (+1) ⚠ |
 | worst CCN | 128 |
 
 ## Documentation
 
 | Metric | Value |
 |---|---:|
-| markdown files | 137 |
-| markdown lines | 28,893 (+31) ⚠ |
+| markdown files | 138 (+1) ⚠ |
+| markdown lines | 28,986 (+93) ⚠ |
 | plan files | 31 |
-| backlog lines | 2,986 |
+| backlog lines | 3,025 (+39) ⚠ |
 | lessons lines | 518 |
-| CLAUDE.md lines | 277 |
+| CLAUDE.md lines | 279 (+2) ⚠ |
 
diff --git a/docs/work/future/backlog-core.md b/docs/work/future/backlog-core.md
index f21a10ce..7fc0f6b8 100644
--- a/docs/work/future/backlog-core.md
+++ b/docs/work/future/backlog-core.md
@@ -832,15 +832,46 @@ Run a bare-leaving scenario before a tree-assuming one and the latter fails pre-
 
 Fix options: (a) make every live mutate scenario clear+rebuild its own canvas (consistent with the newer ones) so order never matters; or (b) have the live runner reboot / restore the canonical tree between scenarios. (a) is the cleaner long-term shape. Until then, the in-process suite is the gate; live full-suite runs need a clean boot per scenario, or run scenarios individually.
 
+## Ethernet on the QuinLED Dig-Next-2 (W5500 over SPI)
+
+The DN2 is in the catalog with no Ethernet block, so it comes up WiFi-only. Adding it is a `deviceModels.json` entry and nothing else: the preset work landed the seam, and W5500 already applies live.
+
+**RMII is impossible on this board, so the W5500 is the only route.** The DN2 carries an ESP32-PICO-V3-02, whose package leaves GPIO 16, 17, 18 and 23 unbonded, and those are the RMII management and clock pins. This is the same package fact behind the flash-cache wedge recorded in [lessons](../past/lessons.md). QuinLED's own LAN8720 board stacks onto the QuinLED-ESP32 only, not onto a Dig board.
+
+**The wiring, from [Quindor's firmware source](https://github.com/intermittech/WLED-Dig-Next-2_W5500)**, using the QEXT header for the bus and the Stemma QT header for the two control lines. Fitting Ethernet therefore costs the I2C port:
+
+| W5500 | connector | GPIO |
+|---|---|---|
+| SCLK | QEXT green | 25 |
+| MOSI | QEXT blue | 32 |
+| MISO | QEXT yellow | 0 |
+| CS | QEXT white | 33 |
+| INT | Stemma QT SDA | 15 |
+| RST | Stemma QT SCL | 14 |
+
+MISO on GPIO 0 is deliberate and sound: GPIO 0 is a strapping pin, and a W5500 leaves MISO high-Z while CS is inactive, which is what Espressif's rule about not driving strapping pins at reset asks for.
+
+**It stays `Custom`, and that is the finding.** Across the three W5500 boards now known, LightCrafter 16, SE 16 V1 and this one, no two share a single pin on any of the six lines. SPI Ethernet is a module wired to whatever a board left free, so there is no W5500 pattern to make a preset from, and inventing one would be a preset that fits nothing.
+
+**What is unverified.** The pin table is trustworthy as what that firmware does, because it was read out of the code rather than a description. What rests on one source is that it works: a single-author repository created 2026-09-19, self-labelled experimental, tested on one unit, with LED output, buttons, relays and mic explicitly not exercised. No independent build was reported when this was written. So the open question is not the pins but whether SPI Ethernet and the LED output coexist on this board under load, which is a bench answer.
+
 ## Housekeeping
 
-### Promote the one-line `//` cap to an error, once the 606 `.cpp` runs are homed (2026-09-21)
+### Promote the one-line `//` cap to an error, once the 604 `.cpp` blocks are homed (2026-09-21)
+
+`MAX_CODE_COMMENT = 1` is already the rule in both kinds of file, and already an ERROR in a header. It stays a warning in a `.cpp` only because `_blocks()` stages the sweep by file kind, and that function names its own removal: a rule leaves the staging as the tree meets it, which no-hard-wrap already did.
+
+**The shape of what is left.** 604 blocks across 92 files, 2081 comment lines. By area: 239 in `test/`, 210 in `src/`, 155 under `src/platform/`. By length: 81 are two lines, 321 are three, 124 are four, so 526 of 604 are three lines or fewer and the tail is one 15-line block, two of ten and a handful between.
+
+It concentrates, which is what makes it approachable. `HttpServerModule.cpp` alone holds 100 and the top ten files hold 312, over half; the other 82 files average under four each.
+
+**Where the depth goes is the blocker.** Of the 92 files only 19 already carry an `@moreinfo` appendix. 68 have a `///` lead with no appendix and 5 have neither, so most files need the appendix written before a single block can move.
 
-`MAX_CODE_COMMENT = 1` is already the rule in both kinds of file, and already an ERROR in a header. It stays a warning in a `.cpp` only because `_blocks()` stages the sweep by file kind, and that function names its own removal: when the implementation side reaches zero it goes and every finding blocks.
+**The fix is not to split them.** Chopping one comment into two shorter lines leaves the text identical and satisfies nothing, which is why the cap counts a block rather than a line. Each one is depth that belongs in the file lead's appendix with an `@xref` back, or prose that says less.
 
-What stands between here and there is 606 multi-line blocks, all `.cpp`: 241 in `test/`, 210 in `src/`, 155 under `src/platform/`. By length, 82 are two lines, 321 are three and 125 are four, so 528 of 606 are three lines or fewer; the tail is one 15-line run, two of ten and a handful between.
+**Suggested order.** `HttpServerModule.cpp` first: 100 blocks, 444 comment lines, and it already has the appendix. Then the other nine heavy files for 212 more. Then `test/` as its own sweep, because the argument that a test comment explains why a test EXISTS is strongest there and may deserve a different budget rather than an appendix. Promote once at zero.
 
-**The fix is not to split them.** Chopping one long comment into two shorter lines leaves the text identical and satisfies nothing, which is why the run cap exists at all. Each one is depth that belongs in the file lead's `@moreinfo` appendix with an `@xref` back, or in prose that says less. That is a judgement per comment across roughly 150 files, which is a sweep of its own rather than a mechanical pass.
+**Cost.** Not a scripted pass. Each block is a judgement between depth that belongs in an appendix, an explanation that should be shorter, and history git already holds. About 35 done by hand in one session is the measured rate, so the first file is a session of its own and the whole sweep is several days.
 
 ### Hot path: move blocking work off the render callbacks (architecture)
 
diff --git a/mkdocs.yml b/mkdocs.yml
index 91603f63..0e6277f7 100644
--- a/mkdocs.yml
+++ b/mkdocs.yml
@@ -144,6 +144,9 @@ nav:
       # Two pages, two audiences: the device card deep-links this one from its consent prompt
       # (app.js, mooncloud.html#why-you-might-like-this), so its URL is a shipped contract.
       - MoonCloud: explanation/mooncloud.md
+      # Answers to what people hit on a first run, each from someone's bench. Under Getting
+      # started rather than How-to, because the reader asking is on their first hour.
+      - FAQ: how-to/faq.md
       - Log an issue: how-to/logging-an-issue.md
   # The FOLDER under docs/ is the Diataxis type: tutorials/, how-to/, explanation/,
   # reference/. The nav labels stay reader-facing, so the path tells a writer where a page
diff --git a/moondeck/MoonDeck.md b/moondeck/MoonDeck.md
index 6b441f4d..d3f78c90 100644
--- a/moondeck/MoonDeck.md
+++ b/moondeck/MoonDeck.md
@@ -184,6 +184,10 @@ And per file, over every `.h` and `.cpp` under `src/`, `test/`, `esp32/main`, `m
 
 Every run writes [docs/reference/metrics/docgen.md](../docs/reference/metrics/docgen.md), the tracked state of the sweep: errors and warnings per rule and per page, then the files ranked within each area. Current state only, so its git history is the trend, the same shape repo-health.md uses. The file is the artifact to read, because stdout scrolls away and truncates; `--noreport` suppresses the write for a caller that only wants the exit code.
 
+**A ratchet, not a snapshot.** The committed `docs/reference/metrics/docgen.md` is the baseline: the run rewrites the file, then compares what it found against the copy in `HEAD`. An error fails the run, and so does any rule whose warning count ROSE, the total included. Warnings are staged work, and staged work that grows is not a sweep.
+
+The comparison is per rule as well as on the total, because each hides a different move. A total alone hides one rule paying for another: splitting an over-wide line lowers the width count and raises the block count, which is the trade the block cap exists to refuse. Per rule alone misses a rule sitting under its own baseline while the total climbs. A rule whose limit itself changed is the one case where a rise is right, and the commit message is where that is said.
+
 **There is no tolerated list.** The limits are the limits, and the check is red until the tree meets them. A grandfather list was tried and removed: while one exists, the cheapest way to make the check green is to add to it, which is how the comment budget eroded in the first place.
 
 The rules are in [documentation-standards.md](../docs/contributing/documentation-standards.md#the-card), and the check itself is pinned by `test/python/test_check_docgen.py`: every rule is tested firing on a page built to break it, because a regex that silently stopped matching would report a clean run.
diff --git a/moondeck/check/check_devices.py b/moondeck/check/check_devices.py
index 385ef03f..0a56d2a5 100644
--- a/moondeck/check/check_devices.py
+++ b/moondeck/check/check_devices.py
@@ -29,6 +29,8 @@
 ROOT = Path(__file__).resolve().parent.parent.parent
 CATALOG = ROOT / "mooninstaller" / "deviceModels.json"
 MAIN_CPP = ROOT / "src" / "main.cpp"
+NETWORK_MODULE = ROOT / "src" / "core" / "system" / "NetworkModule.h"
+PLATFORM_CONFIG = ROOT / "src" / "platform" / "esp32" / "platform_config.h"
 DOCS = ROOT / "docs"
 
 # The device stores the injected deviceModel in SystemModule's deviceModel_[32] buffer (31 usable
@@ -56,6 +58,60 @@
 BOOT_WIRED_TYPES = {"System", "Network", "Drivers"}
 
 
+def eth_preset_labels():
+    """The Ethernet preset labels from NetworkModule.h's kEthPresets table.
+
+    Read from the firmware rather than restated here, so a preset renamed in one place and not the
+    other fails this check instead of silently leaving a catalog entry pointing at nothing.
+    """
+    text = NETWORK_MODULE.read_text(encoding="utf-8")
+    table = re.search(r"kEthPresets\[\]\s*=\s*\{(.*?)\n    \};", text, re.S)
+    if not table:
+        return set()
+    return set(re.findall(r'^\s*\{"([^"]+)"', table.group(1), re.M))
+
+
+def eth_preset_drift():
+    """Preset rows that no longer match the per-chip default they restate, as (label, field, was, now).
+
+    Three presets ARE a chip's `ethConfigDefault`, so the same pin map has two homes in two layers.
+    They cannot simply reference each other: the header's value is a compile-time `constexpr` chosen
+    per chip, while the table must carry every chip's row on every build. So the duplication stands
+    and this check binds it, because the drift is silent where it matters most: `seedEthPresetFromPins`
+    matches on exact equality, so one corrected pin would reseed every provisioned board to Custom.
+    """
+    table = NETWORK_MODULE.read_text(encoding="utf-8")
+    header = PLATFORM_CONFIG.read_text(encoding="utf-8")
+    block = re.search(r"kEthPresets\[\]\s*=\s*\{(.*?)\n    \};", table, re.S)
+    if not block:
+        return [("kEthPresets", "table", "present", "unreadable")]
+
+    rows = {}
+    for m in re.finditer(r'\{"([^"]+)",\s*([^}]+)\}', block.group(1)):
+        rows[m.group(1)] = [f.strip() for f in m.group(2).split(",")]
+
+    # label -> the PHY constant naming that chip's arm of the ethConfigDefault ternary.
+    PAIRS = {"Classic RMII": "ethLan8720", "P4-NANO": "ethIp101", "S31 CoreBoard": "ethYt8531"}
+    FIELDS = ["phyAddr", "mdc", "mdio", "rst", "rmiiClk"]
+    drift = []
+    for label, phy in PAIRS.items():
+        row = rows.get(label)
+        if not row:
+            drift.append((label, "row", "present", "missing"))
+            continue
+        m = re.search(phy + r", /\*addr\*/ (-?\d+), /\*mdc\*/ (-?\d+), /\*mdio\*/ (-?\d+),"
+                      r"\s*/\*rst\*/ (-?\d+), /\*rmiiClk\*/ (-?\d+)", header)
+        if not m:
+            drift.append((label, phy, "an ethConfigDefault arm", "not found"))
+            continue
+        # row is label-less here: [type, phyAddr, mdc, mdio, rst, rmiiClk, ...]
+        for i, field in enumerate(FIELDS):
+            want, got = m.group(i + 1), row[i + 1]
+            if want != got:
+                drift.append((label, field, want, got))
+    return drift
+
+
 def registered_types():
     """The set of factory type names from main.cpp's registerType<T>("Name") calls."""
     text = MAIN_CPP.read_text(encoding="utf-8")
@@ -64,6 +120,13 @@ def registered_types():
 
 def main():
     errors = []
+    eth_presets = eth_preset_labels()
+    for label, field, was, now in eth_preset_drift():
+        errors.append(f"NetworkModule.h kEthPresets {label!r}: {field} is {now}, but "
+                      f"platform_config.h's ethConfigDefault says {was} — the preset restates the "
+                      f"chip default, so the two must agree or a provisioned board reseeds to Custom")
+    if not eth_presets:
+        errors.append("NetworkModule.h: could not read the kEthPresets table — the ethBoard check cannot run")
 
     try:
         catalog = json.loads(CATALOG.read_text(encoding="utf-8"))
@@ -247,6 +310,12 @@ def _gpio(v):
             # are genuine BOARD WIRING are required — MDC/MDIO may stay at the IDF default (omit or
             # -1) on RMII, since that's a real standard, not a board-specific value.
             if mtype == "NetworkModule" and isinstance(controls, dict):
+                board = controls.get("ethBoard")
+                if board is not None:
+                    if not isinstance(board, str):
+                        errors.append(f"{where}: NetworkModule ethBoard must be the preset LABEL as a string, got {board!r}")
+                    elif board not in eth_presets:
+                        errors.append(f"{where}: NetworkModule ethBoard {board!r} is not a preset in NetworkModule.h (known: {sorted(eth_presets)})")
                 et = controls.get("ethType")
                 # ethType must be an int (a JSON string like "2" would silently skip the rule below and
                 # also isn't what the device deserializes into the Select) — reject a stringified value.
@@ -254,7 +323,9 @@ def _gpio(v):
                 # would pass as ethType 1 (LAN8720).
                 if et is not None and type(et) is not int:
                     errors.append(f"{where}: NetworkModule ethType must be an integer, got {et!r}")
-                if type(et) is int and et != 0:
+                # A named preset IS the explicit pinning: the map lives in NetworkModule.h, under
+                # the same review as any other firmware constant, rather than repeated per board.
+                if type(et) is int and et != 0 and board in (None, "Custom"):
                     # RMII LAN8720(1)/IP101(2): rst + clock. RGMII YT8531(4): mdc/mdio/rst.
                     # W5500 SPI(3): the four SPI bus pins. -1 is an allowed explicit value ("unused /
                     # IDF default"); what the rule forbids is OMITTING a board-wiring pin.
diff --git a/moondeck/check/check_docgen.py b/moondeck/check/check_docgen.py
index ee236813..4486b95b 100755
--- a/moondeck/check/check_docgen.py
+++ b/moondeck/check/check_docgen.py
@@ -147,6 +147,10 @@
 # (`lead N chars > M`), and the prefix alone also matched the word budget (`comment line N words`),
 # each a differently motivated rule silently demoted.
 _LINE_LENGTH_RULES = ("doc line ", "comment line ")
+# The one rule staged the OTHER way: an implementation file's hard wrap blocks too. The tree is at
+# zero, so nothing is held hostage, and a split sentence reflows every line it spans on the next
+# word change, which costs a reviewer the same in a `.cpp` as in a header.
+_HARD_WRAP_RULE = "hard wrap"
 # One line, in every file. Depth is not forbidden, it is HOMED: a header carries it in an
 # `@moreinfo` appendix (212 do) and an implementation file carries it in the same appendix on its
 # own file lead. A cap of 8 in a `.cpp` was an attempt to give depth a home inline, which leaves
@@ -182,12 +186,16 @@ def _blocks(key: str, why: str = "") -> bool:
     goes and every finding blocks, the same way `.vale.ini` promotes a page to error once the
     sweep has finished it and loses its per-page section when the last page lands.
 
-    ONE rule is staged separately: the line-length cap warns in a header too. Every other rule
-    describes how this tree is already written, so a finding is a defect against a standard the
-    file was written under. The cap is new, so its findings are lines nobody wrote wrongly, and
-    erroring on 820 of them would stop commits over a target rather than a defect. It joins the
-    others as an error once the tree meets it, which is the same staging this function is.
+    ONE rule is staged in each direction. The line-length cap warns in a header too: every other
+    rule describes how this tree is already written, so a finding is a defect against a standard
+    the file was written under, while the cap is new and erroring on 820 lines nobody wrote
+    wrongly would stop commits over a target. And NO-HARD-WRAP blocks in both kinds of file: the
+    tree meets it, at zero findings, so the staging has nothing left to stage. The reason holds
+    everywhere too, being the diff rather than the page, and a rule met everywhere is a rule that
+    should hold everywhere.
     """
+    if why.startswith(_HARD_WRAP_RULE):
+        return True
     if why.startswith(_LINE_LENGTH_RULES) and "chars >" in why:
         return False
     return _generates_a_page(key.partition("::")[0])
@@ -1570,6 +1578,57 @@ def _report(entries, heading: str) -> None:
             print(f"    {title}: {why}")
 
 
+def _committed_counts():
+    """Every rule's warning count in the COMMITTED report, or None when there is no baseline.
+
+    The report is tracked, so the last commit's copy is the number to beat. Read from git rather
+    than from the working tree, because the run rewrites the file before the comparison happens.
+    """
+    import subprocess
+    rel = REPORT.relative_to(ROOT).as_posix()
+    r = subprocess.run(["git", "show", f"HEAD:{rel}"],
+                       cwd=ROOT, capture_output=True, text=True)
+    if r.returncode != 0:
+        return None
+    counts = {}
+    # Only the `## By rule` table: the area and file tables share its column shape, and a row from
+    # either would enter the baseline as a rule that can never appear in `now`.
+    in_rules = False
+    for line in r.stdout.split("\n"):
+        if line.startswith("## "):
+            in_rules = line.strip() == "## By rule"
+        m = re.match(r"^\| (.+?) \| (\d+) \| (\d+) \|$", line.strip())
+        if in_rules and m and m.group(1) != "Rule":
+            counts[m.group(1)] = int(m.group(3))
+        # The TOTAL as its own entry, because a per-rule comparison alone misses a rule that is
+        # below the baseline and still rising: one under budget can absorb a new finding silently.
+        m = re.match(r"^\*\*\d+ error\(s\)\*\* and \*\*(\d+) warning\(s\)\*\* across", line.strip())
+        if m and "(total)" not in counts:
+            counts["(total)"] = int(m.group(1))
+    return counts or None
+
+
+def _ratchet(found) -> list:
+    """Which rules rose against the committed report. The list only shrinks.
+
+    A warning is staged work, and staged work that grows is not a sweep. Per RULE rather than on
+    the total alone, because a total hides one rule paying for another: splitting a long line
+    lowers the width count and raises the block count, which is the move the run cap exists to
+    refuse.
+    """
+    from collections import Counter
+    base = _committed_counts()
+    if base is None:
+        return []
+    # LABELLED, matching how the report writes them: `_rule_name` returns the internal key
+    # ("code comment") while the table carries the display label ("multi-line comment blocks"),
+    # and comparing the two key spaces silently compares nothing.
+    now = Counter(_rule_label(_rule_name(why)) for key, why in found if not _blocks(key, why))
+    now["(total)"] = sum(1 for key, why in found if not _blocks(key, why))
+    return [(rule, was, now.get(rule, 0))
+            for rule, was in base.items() if now.get(rule, 0) > was]
+
+
 def main() -> int:
     ap = argparse.ArgumentParser(description=__doc__)
     ap.add_argument("--noreport", action="store_true",
@@ -1595,7 +1654,13 @@ def main() -> int:
               f"{MAX_DOC_WORDS} words.")
         return 0
 
+    risen = _ratchet(found)
     print(f"Docgen check: {len(errors)} error(s), {len(warnings)} warning(s).\n")
+    if risen:
+        print("WARNINGS ROSE against the committed report. The list only shrinks:")
+        for rule, was, now in risen:
+            print(f"  {rule}: {was} -> {now}")
+        print("  Fix them, or say in the commit why the rule itself changed.\n")
     if errors:
         _report(errors, "ERRORS, in files that generate a page. These fail the gate.")
     if warnings:
@@ -1607,7 +1672,7 @@ def main() -> int:
           "\n(the technical page the card links), cross-module rationale into a"
           "\n`## <Name>, details` section on the same page."
           "\nRules: docs/contributing/documentation-standards.md § The card.")
-    return 1 if errors else 0
+    return 1 if (errors or risen) else 0
 
 
 if __name__ == "__main__":
diff --git a/moondeck/check/check_prose.py b/moondeck/check/check_prose.py
index c7036fbf..ac16ff99 100755
--- a/moondeck/check/check_prose.py
+++ b/moondeck/check/check_prose.py
@@ -126,7 +126,7 @@ def main():
     # the tree. A zero is only trustworthy once something that should fire, fires.
     probe = subprocess.run(
         ["vale", "--output=JSON", "--no-exit", "--ext=.h", "--path=probe.h"],
-        input="// A color scheme \u2014 an em-dash.\nint x = 1;\n",
+        input="// A colour scheme \u2014 an em-dash.\nint x = 1;\n",
         capture_output=True, text=True)
     try:
         hits = sum(len(v) for v in json.loads(probe.stdout or "{}").values())
diff --git a/mooninstaller/deviceModels.json b/mooninstaller/deviceModels.json
index 20743691..a69c8c33 100644
--- a/mooninstaller/deviceModels.json
+++ b/mooninstaller/deviceModels.json
@@ -35,11 +35,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 1,
-          "ethPhyAddr": 0,
-          "ethRstGpio": 5,
-          "ethClockGpio": 17,
-          "ethClockExtIn": false
+          "ethBoard": "Classic RMII"
         }
       }
     ]
@@ -329,13 +325,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 1,
-          "ethPhyAddr": 0,
-          "ethClockGpio": 17,
-          "ethClockExtIn": false,
-          "ethMdcGpio": 23,
-          "ethMdioGpio": 18,
-          "ethRstGpio": -1
+          "ethBoard": "Classic RMII (no reset)"
         }
       },
       {
@@ -782,13 +772,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 2,
-          "ethPhyAddr": 1,
-          "ethMdcGpio": 31,
-          "ethMdioGpio": 52,
-          "ethRstGpio": 51,
-          "ethClockGpio": 50,
-          "ethClockExtIn": true
+          "ethBoard": "P4-NANO"
         }
       }
     ]
@@ -840,11 +824,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 1,
-          "ethPhyAddr": 0,
-          "ethRstGpio": 5,
-          "ethClockGpio": 17,
-          "ethClockExtIn": false
+          "ethBoard": "Classic RMII"
         }
       },
       {
@@ -898,6 +878,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
+          "ethBoard": "Custom",
           "ethType": 3,
           "ethSpiMiso": 13,
           "ethSpiMosi": 11,
@@ -957,6 +938,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
+          "ethBoard": "Custom",
           "ethType": 3,
           "ethSpiMiso": 5,
           "ethSpiMosi": 6,
@@ -1026,13 +1008,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 2,
-          "ethPhyAddr": 1,
-          "ethMdcGpio": 31,
-          "ethMdioGpio": 52,
-          "ethRstGpio": 51,
-          "ethClockGpio": 50,
-          "ethClockExtIn": true
+          "ethBoard": "P4-NANO"
         }
       },
       {
@@ -1098,13 +1074,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 2,
-          "ethPhyAddr": 1,
-          "ethMdcGpio": 31,
-          "ethMdioGpio": 52,
-          "ethRstGpio": 51,
-          "ethClockGpio": 50,
-          "ethClockExtIn": true
+          "ethBoard": "P4-NANO"
         }
       }
     ]
@@ -1153,11 +1123,7 @@
         "type": "NetworkModule",
         "id": "Network",
         "controls": {
-          "ethType": 4,
-          "ethPhyAddr": -1,
-          "ethMdcGpio": 5,
-          "ethMdioGpio": 6,
-          "ethRstGpio": 7
+          "ethBoard": "S31 CoreBoard"
         }
       },
       {
diff --git a/mooninstaller/install.js b/mooninstaller/install.js
index 48b7cf2f..9fbde554 100644
--- a/mooninstaller/install.js
+++ b/mooninstaller/install.js
@@ -1150,7 +1150,13 @@ document.addEventListener('DOMContentLoaded', () => {
       // active wherever NetworkModule exists (the radio is always available); a board
       // that lists WiFi as supported but ships no NetworkModule entry stays "supported".
       const ethConfigured = (m) => {
-        const t = m.controls && m.controls.ethType;
+        const c = m.controls;
+        if (!c) return false;
+        // A named board preset IS the configuration: it carries the PHY and the pins, so a
+        // board that picks one lists no ethType of its own. Custom is the escape hatch and
+        // names no PHY, so it falls through to the ethType its entry must then carry.
+        if (c.ethBoard !== undefined && c.ethBoard !== "Custom") return true;
+        const t = c.ethType;
         return t !== undefined && t !== 0 && t !== "0" && t !== "None";
       };
       const CAP_MODULE = {
diff --git a/src/core/module/Control.h b/src/core/module/Control.h
index d31403e1..6c65d83d 100644
--- a/src/core/module/Control.h
+++ b/src/core/module/Control.h
@@ -138,6 +138,11 @@ struct ListSource {
                                  const char* /*valueJson*/) { return false; }
 };
 
+// How much a reader wants to see. One number the whole UI composes against, so a control names the audience it is for rather than every card deciding for itself.
+inline constexpr uint8_t kModeUser = 0;       ///< the show: what a light does
+inline constexpr uint8_t kModeExpert = 1;     ///< and the tuning an installation needs: the peripheral, the pin, the rate
+inline constexpr uint8_t kModeDeveloper = 2;  ///< and what diagnoses the firmware, meaning nothing without the source beside it
+
 /// One control's metadata: what it points at, how to render it, and how to persist it.
 ///
 /// The value lives in the module's own variable, and this borrows a pointer to it.
@@ -159,8 +164,8 @@ struct ControlDescriptor {
     bool persistLabel = false;
     /// Whether the UI renders this display-only, for a value tooling pushes rather than a user.
     bool readonly = false;
-    /// Whether this is expert-only, shown by the UI only while expert mode is on.
-    bool advanced = false;
+    /// The mode a reader needs before this control is shown: 0 everyone, 1 expert, 2 developer.
+    uint8_t minMode = 0;
     /// Whether a numeric renders as a number input, for an integer that is an address not a magnitude.
     bool numberField = false;
     // These sit after the other flags, since the text initializers below are positional.
@@ -370,9 +375,14 @@ class ControlList {
         if (i < count_) controls_[i].readonly = readonly;
     }
 
-    /// Mark a control expert-only, which the UI shows only while expert mode is on.
+    /// Mark a control expert-only, which the UI shows from expert mode up.
     void setAdvanced(uint8_t i, bool advanced = true) {
-        if (i < count_) controls_[i].advanced = advanced;
+        if (i < count_) controls_[i].minMode = advanced ? kModeExpert : kModeUser;
+    }
+
+    /// Mark a control developer-only: a number that diagnoses the firmware rather than the show.
+    void setDeveloper(uint8_t i) {
+        if (i < count_) controls_[i].minMode = kModeDeveloper;
     }
 
     /// Render a numeric as a number input, for an integer that is an identity not a magnitude.
diff --git a/src/core/module/MoonModule.h b/src/core/module/MoonModule.h
index 53889d5f..48983e31 100644
--- a/src/core/module/MoonModule.h
+++ b/src/core/module/MoonModule.h
@@ -143,7 +143,7 @@ class MoonModule {
             mix(static_cast<uint32_t>(c.type));
             mix(static_cast<uint32_t>(c.min));
             mix(static_cast<uint32_t>(c.max));
-            mix((c.hidden ? 1u : 0u) | (c.readonly ? 2u : 0u) | (c.advanced ? 4u : 0u));
+            mix((c.hidden ? 1u : 0u) | (c.readonly ? 2u : 0u) | (static_cast<unsigned>(c.minMode) << 2));
             if (c.type == ControlType::Select && c.aux) {
                 // Hash the strings, so an in-place rename still changes the signature.
                 const char* const* opts = reinterpret_cast<const char* const*>(c.aux);
diff --git a/src/core/system/HttpServerModule.cpp b/src/core/system/HttpServerModule.cpp
index 72db749f..6794d582 100644
--- a/src/core/system/HttpServerModule.cpp
+++ b/src/core/system/HttpServerModule.cpp
@@ -1220,13 +1220,16 @@ void HttpServerModule::writeModuleJson(JsonSink& sink, MoonModule* mod) {
     const char* roleStr = roleName(mod->role());
     const char* type = mod->typeName();
     if (!type) type = "";
+    // `enabled` is published for every module, and the ones that keep running regardless say so here, so the UI can leave out a switch that would do nothing.
     sink.appendf(
         "{\"name\":\"%s\",\"type\":\"%s\",\"role\":\"%s\",\"enabled\":%s,"
+        "\"respectsEnabled\":%s,"
         "\"tickTimeUs\":%u,\"classSize\":%u,\"dynamicBytes\":%u",
         mod->name() ? mod->name() : "",
         type,
         roleStr,
         mod->enabled() ? "true" : "false",
+        mod->respectsEnabled() ? "true" : "false",
         static_cast<unsigned>(mod->tickTimeUs()),
         static_cast<unsigned>(mod->classSize()),
         static_cast<unsigned>(mod->dynamicBytes()));
@@ -1304,7 +1307,7 @@ void HttpServerModule::writeControls(JsonSink& sink, MoonModule* mod) {
         writeControlMetadata(sink, c);
         // Emit optional flags only when set (common case is false; omit to save bytes).
         if (c.readonly) sink.append(",\"readonly\":true");
-        if (c.advanced) sink.append(",\"advanced\":true");   // UI shows it only in expert mode
+        if (c.minMode) sink.appendf(",\"minMode\":%u", static_cast<unsigned>(c.minMode));   // the mode the UI needs before it shows this
         if (c.numberField) sink.append(",\"numberField\":true");   // render a plain number input, not a slider
         // An editable List (the CRUD primitive) tells the UI to show add/delete/reorder + inline row editors; a plain List stays read-only. The row objects carry a stable "id" the /api/list/* ops address, and each editable row's detail carries its field descriptors.
         if (c.switchRow) sink.append(",\"switchRow\":true");
diff --git a/src/core/system/NetworkModule.h b/src/core/system/NetworkModule.h
index 068a804d..ba6f91d7 100644
--- a/src/core/system/NetworkModule.h
+++ b/src/core/system/NetworkModule.h
@@ -186,9 +186,24 @@ class NetworkModule : public MoonModule {
         controls_.addIPv4("dns", staticDns_);
         controls_.setHidden(controls_.count() - 1, hideStatic);
 
-        // Only where a driver is compiled in, the type selecting which pin rows apply.
-        if constexpr (platform::hasEthernet) {
+        // Where a driver is compiled in, the type selecting which pin rows apply; a desktop builds the same rows with no interface behind them, to exercise the presets.
+        if constexpr (platform::hasEthernet || platform::previewsEthernetControls) {
+            // A preview configures nothing, so it is developer-mode only and says so on the card.
+            constexpr bool preview = !platform::hasEthernet;
+            const uint8_t firstEthControl = controls_.count();
+            buildEthPresetOptions();
+            // First render: on a provisioned board the pins are already the catalog's, so read the preset back off them. A factory board has no type yet and lands on Custom with its chip's pins, which is honest until the catalog names a board.
+            if (!ethPresetSeeded_) { ethPresetSeeded_ = true; seedEthPresetFromPins(); }
+            // The restore path fires no onControlChanged, so writing the map HERE is what makes a saved `ethBoard` reach the pins. Keyed on the selection moving, or a rebuild would undo a Custom edit.
+            else if (ethPresetSel_ != ethPresetApplied_) applyEthPreset();
+            ethPresetApplied_ = ethPresetSel_;
+            controls_.addSelect("ethBoard", ethPresetSel_, ethPresetOptions_, ethPresetCount_);
+            // By label: the list is filtered per build, so an index would name a different board.
+            controls_.setPersistLabel(controls_.count() - 1);
+            // Hidden on a known board, and hidden stays BOUND: the values still drive the interface.
+            const bool editable = ethPinsEditable();
             controls_.addSelect("ethType", ethType_, ethTypeOptions_, 5);
+            controls_.setHidden(controls_.count() - 1, !editable);
             const bool isRmii  = (ethType_ == 1 || ethType_ == 2);
             const bool isSpi   = (ethType_ == 3);
             const bool isRgmii = (ethType_ == 4);
@@ -197,29 +212,33 @@ class NetworkModule : public MoonModule {
             // An address rather than a GPIO, and signed for the auto-detect sentinel.
             controls_.addControl("ethPhyAddr", ethPhyAddr_, -1, 31);
             controls_.setNumberField(controls_.count() - 1);   // an identity, not a magnitude
-            controls_.setHidden(controls_.count() - 1, !isEth);
+            controls_.setHidden(controls_.count() - 1, !editable || !isEth);
             controls_.addPin("ethRstGpio", ethRstGpio_);
-            controls_.setHidden(controls_.count() - 1, !isEth);
+            controls_.setHidden(controls_.count() - 1, !editable || !isEth);
             // Every wired interface needs them, and showing them is what the pin map counts.
             controls_.addPin("ethMdcGpio", ethMdcGpio_);
-            controls_.setHidden(controls_.count() - 1, !isRmii && !isRgmii);
+            controls_.setHidden(controls_.count() - 1, !editable || (!isRmii && !isRgmii));
             controls_.addPin("ethMdioGpio", ethMdioGpio_);
-            controls_.setHidden(controls_.count() - 1, !isRmii && !isRgmii);
+            controls_.setHidden(controls_.count() - 1, !editable || (!isRmii && !isRgmii));
             controls_.addPin("ethClockGpio", ethClockGpio_);
-            controls_.setHidden(controls_.count() - 1, !isRmii);
+            controls_.setHidden(controls_.count() - 1, !editable || !isRmii);
             // A direction, so a toggle rather than a range.
             controls_.addControl("ethClockExtIn", ethClockExtIn_);
-            controls_.setHidden(controls_.count() - 1, !isRmii);
+            controls_.setHidden(controls_.count() - 1, !editable || !isRmii);
             controls_.addPin("ethSpiMiso", ethSpiMiso_);
-            controls_.setHidden(controls_.count() - 1, !isSpi);
+            controls_.setHidden(controls_.count() - 1, !editable || !isSpi);
             controls_.addPin("ethSpiMosi", ethSpiMosi_);
-            controls_.setHidden(controls_.count() - 1, !isSpi);
+            controls_.setHidden(controls_.count() - 1, !editable || !isSpi);
             controls_.addPin("ethSpiSck", ethSpiSck_);
-            controls_.setHidden(controls_.count() - 1, !isSpi);
+            controls_.setHidden(controls_.count() - 1, !editable || !isSpi);
             controls_.addPin("ethSpiCs", ethSpiCs_);
-            controls_.setHidden(controls_.count() - 1, !isSpi);
+            controls_.setHidden(controls_.count() - 1, !editable || !isSpi);
             controls_.addPin("ethSpiIrq", ethSpiIrq_);
-            controls_.setHidden(controls_.count() - 1, !isSpi);
+            controls_.setHidden(controls_.count() - 1, !editable || !isSpi);
+            // One loop rather than a call beside every row: the whole group carries one tag.
+            if constexpr (preview) {
+                for (uint8_t i = firstEthControl; i < controls_.count(); i++) controls_.setDeveloper(i);
+            }
         }
     }
 
@@ -494,6 +513,116 @@ class NetworkModule : public MoonModule {
     uint32_t appliedAddressingSig_ = 0;
     bool addressingSigApplied_ = false;
 
+    // One board's Ethernet wiring, the fields matching EthPinConfig; -1 leaves a line unused.
+    struct EthPreset {
+        const char* label;
+        int8_t type;          // an EthPhyType
+        int8_t phyAddr;
+        int8_t mdc, mdio, rst;
+        int8_t rmiiClock;
+        bool   rmiiClockExtIn;
+        int8_t miso, mosi, sck, cs, irq;
+        bool   editable;      // false on a soldered map: those lines are not the user's to set
+    };
+
+    // The presets this module knows, each a board family rather than one product; Custom keeps whatever is in the fields, for a hand-wired board.
+    static constexpr EthPreset kEthPresets[] = {
+        // The LAN8720 reference wiring most classic boards follow, which is also the chip default.
+        {"Classic RMII",            1,  0, 23, 18,  5, 17, false, -1, -1, -1, -1, -1, false},
+        // The same wiring with no reset, for a board using GPIO 5 as an LED lane: the PHY resets by jumper.
+        {"Classic RMII (no reset)", 1,  0, 23, 18, -1, 17, false, -1, -1, -1, -1, -1, false},
+        // Waveshare P4-NANO and the boards following its shield pinout: IP101, the clock fed in.
+        {"P4-NANO",           2,  1, 31, 52, 51, 50, true,  -1, -1, -1, -1, -1, false},
+        // The S31's 1 Gb PHY, addressed by scan rather than by a fixed address.
+        {"S31 CoreBoard",     4, -1,  5,  6,  7, -1, false, -1, -1, -1, -1, -1, false},
+        {"Custom",            0, -1, -1, -1, -1, -1, false, -1, -1, -1, -1, -1, true},
+    };
+    static constexpr uint8_t kEthPresetCount = sizeof(kEthPresets) / sizeof(kEthPresets[0]);
+
+    // Which board wiring is selected, the pins following from it unless it is Custom.
+    uint8_t ethPresetSel_ = 0;
+    const char* ethPresetOptions_[kEthPresetCount] = {};
+    uint8_t ethPresetIndex_[kEthPresetCount] = {};
+    uint8_t ethPresetCount_ = 0;
+    bool ethPresetSeeded_ = false;   ///< the preset is matched to the pins once, not every rebuild
+    uint8_t ethPresetApplied_ = 0;   ///< the selection whose map is already written, so a rebuild is not a re-apply
+
+
+    // A preset naming a PHY this build cannot drive would offer pins that reach nothing.
+    /// Does this firmware carry a driver for the preset's PHY?
+    static bool presetBuildable(const EthPreset& p) {
+        if (p.type == 0) return true;                       // Custom, which names no PHY
+        if (p.type == 3) return platform::hasEthW5500;      // SPI, a separate driver
+        return !platform::hasEthW5500;                      // the internal EMAC drives the rest
+    }
+
+    /// Offer the presets this build can drive, re-pointing the selection by label.
+    void buildEthPresetOptions() {
+        const char* current = (ethPresetSel_ < ethPresetCount_) ? ethPresetOptions_[ethPresetSel_] : nullptr;
+        ethPresetCount_ = 0;
+        for (uint8_t i = 0; i < kEthPresetCount; i++) {
+            if (!presetBuildable(kEthPresets[i])) continue;
+            ethPresetOptions_[ethPresetCount_] = kEthPresets[i].label;
+            ethPresetIndex_[ethPresetCount_] = i;
+            ethPresetCount_++;
+        }
+        // By LABEL, so a filtered list cannot silently select a different board.
+        uint8_t sel = 0;
+        if (current) {
+            for (uint8_t k = 0; k < ethPresetCount_; k++) {
+                if (std::strcmp(ethPresetOptions_[k], current) == 0) { sel = k; break; }
+            }
+        }
+        ethPresetSel_ = sel;
+    }
+
+    /// Are the pin controls the user's to edit, for the preset currently selected?
+    bool ethPinsEditable() const {
+        if (ethPresetSel_ >= ethPresetCount_) return true;   // nothing resolved: never hide
+        return kEthPresets[ethPresetIndex_[ethPresetSel_]].editable;
+    }
+
+    // Custom writes nothing, so switching to it after an edit keeps the edit.
+    /// Write the chosen preset's map into the pin controls.
+    void applyEthPreset() {
+        if (ethPresetSel_ >= ethPresetCount_) return;
+        const EthPreset& p = kEthPresets[ethPresetIndex_[ethPresetSel_]];
+        if (p.editable) return;
+        ethType_       = static_cast<uint8_t>(p.type);
+        ethPhyAddr_    = p.phyAddr;
+        ethMdcGpio_    = p.mdc;
+        ethMdioGpio_   = p.mdio;
+        ethRstGpio_    = p.rst;
+        ethClockGpio_  = p.rmiiClock;
+        ethClockExtIn_ = p.rmiiClockExtIn;
+        ethSpiMiso_    = p.miso;
+        ethSpiMosi_    = p.mosi;
+        ethSpiSck_     = p.sck;
+        ethSpiCs_      = p.cs;
+        ethSpiIrq_     = p.irq;
+    }
+
+    // The board a set of pins came from, so a provisioned device opens on its own name.
+    /// The preset whose map these pin values already are, or Custom when none matches.
+    void seedEthPresetFromPins() {
+        for (uint8_t k = 0; k < ethPresetCount_; k++) {
+            const EthPreset& p = kEthPresets[ethPresetIndex_[k]];
+            if (p.editable) continue;
+            if (p.type == static_cast<int8_t>(ethType_) && p.phyAddr == ethPhyAddr_ &&
+                p.mdc == ethMdcGpio_ && p.mdio == ethMdioGpio_ && p.rst == ethRstGpio_ &&
+                p.rmiiClock == ethClockGpio_ && p.rmiiClockExtIn == ethClockExtIn_ &&
+                p.miso == ethSpiMiso_ && p.mosi == ethSpiMosi_ && p.sck == ethSpiSck_ &&
+                p.cs == ethSpiCs_ && p.irq == ethSpiIrq_) {
+                ethPresetSel_ = k;
+                return;
+            }
+        }
+        // Custom is the last row, and the only editable one.
+        for (uint8_t k = 0; k < ethPresetCount_; k++) {
+            if (kEthPresets[ethPresetIndex_[k]].editable) { ethPresetSel_ = k; return; }
+        }
+    }
+
     /// A cheap hash over the interface controls, so a live change is detected.
     uint32_t ethSig() const {
         uint32_t h = ethType_;
diff --git a/src/core/system/SystemModule.h b/src/core/system/SystemModule.h
index d9639387..066f46e8 100644
--- a/src/core/system/SystemModule.h
+++ b/src/core/system/SystemModule.h
@@ -137,11 +137,11 @@ class SystemModule : public MoonModule {
         controls_.addReadOnly("cpu", const_cast<char*>(platform::cpuInfo()));   ///< the clock and core count
         controls_.addReadOnly("sdk", const_cast<char*>(platform::sdkVersion()));
         controls_.addReadOnly("bootReason", const_cast<char*>(platform::resetReason()));
-        // The UI honors this client-side, so nothing in the firmware reads it.
-        controls_.addControl("expertMode", expertMode_);
+        // The UI honors this client-side, so nothing in the firmware reads it. A LEVEL rather than a switch, because a control names the audience it is for and one number decides which audience is reading.
+        controls_.addSelect("mode", mode_, kModeOptions, 3);
         // Warn keeps the once-a-second line off the wire while warnings still print.
         controls_.addSelect("logLevel", logLevel_, logLevelOptions_, 6);
-        controls_.setAdvanced(controls_.count() - 1);
+        controls_.setDeveloper(controls_.count() - 1);
         // Compiled out where the radio is native, so the control and its query both vanish.
         if constexpr (platform::hasWifiCoprocessor) {
             controls_.addReadOnly("wifiCoproc", const_cast<char*>(platform::coprocessorWifi()));
@@ -217,7 +217,8 @@ class SystemModule : public MoonModule {
     Scheduler* scheduler_ = nullptr;
 
     char deviceName_[24] = {};   ///< the one network identity
-    bool expertMode_ = false;    ///< one flag the whole UI composes against
+    uint8_t mode_ = 0;           ///< one level the whole UI composes against: user, expert, developer
+    static constexpr const char* kModeOptions[3] = {"user", "expert", "developer"};
     /// Push the level to the logger, clamped so a corrupt value cannot index past the end.
     void applyLogLevel() {
         uint8_t lvl = logLevel_ > static_cast<uint8_t>(platform::LogLevel::Verbose)
diff --git a/src/platform/desktop/platform_config.h b/src/platform/desktop/platform_config.h
index d6f19128..7c43ae3a 100644
--- a/src/platform/desktop/platform_config.h
+++ b/src/platform/desktop/platform_config.h
@@ -94,6 +94,9 @@ struct EthPinConfig {
 /// False, so shared code compiles its Ethernet controls out and seeds itself from the default below.
 constexpr bool hasEthernet = false;
 
+/// True: the Ethernet CONTROLS are built though no interface exists, so the board presets can be exercised without a device. A preview of a device-only card, tagged developer-mode and never consulted by a bring-up path, which stays keyed on hasEthernet.
+constexpr bool previewsEthernetControls = true;
+
 /// True: a host has several NICs, so a raw sender must name the one it binds.
 constexpr bool hasNamedNetInterfaces = true;
 
diff --git a/src/platform/esp32/platform_config.h b/src/platform/esp32/platform_config.h
index 08e6e945..068c8239 100644
--- a/src/platform/esp32/platform_config.h
+++ b/src/platform/esp32/platform_config.h
@@ -188,6 +188,9 @@ constexpr bool hasEthernet = true;
 // True when the firmware carries an IP stack at all, which is what UDP interop gates on.
 constexpr bool hasNetwork = hasWiFi || hasEthernet;
 
+/// False: the controls follow the hardware, so a board without Ethernet shows nothing to configure.
+constexpr bool previewsEthernetControls = false;
+
 // True only under emulation, where a saved type would otherwise select hardware that is not there.
 #ifdef CONFIG_ETH_USE_OPENETH
 constexpr bool ethPhyIsFixed = true;
diff --git a/src/ui/app.js b/src/ui/app.js
index e019555c..c78e5d6c 100644
--- a/src/ui/app.js
+++ b/src/ui/app.js
@@ -55,12 +55,10 @@ const dragTs = {};               // per-control last-touched timestamp (ms): a s
 const EDITABLE_CONTROL_TYPES = new Set(
     ["uint8", "uint16", "int16", "int32", "pin", "bool", "text", "textarea", "filepath", "password", "select",
      "palette", "ipv4"]);
-const TIMING_MODES = ["fps", "ms"];
 
 // localStorage keys per ui.md
 const LS_SELECTED  = "mm_selectedRoot";
 const LS_THEME     = "mm_theme";
-const LS_TIMING    = "mm_timing_mode";
 const LS_TABS      = "mm_selectedTabs";   // { [containerName]: childName }: the open tab per container
 const LS_EXPANDED  = "mm_expanded";       // [moduleName, …]: modules whose "controls" <details> is open
 const LS_TA_SIZE   = "mm_textareaSizes";  // { "<module>:<control>": heightPx }: user-dragged textarea heights
@@ -149,7 +147,7 @@ function lsRead(key, defaultVal) {
 // this module were the only thing running. On a card that is a claim about the whole pipeline made
 // from one part of it, so the honest default is the microseconds. The toggle still cycles to fps
 // for anyone comparing against a target rate.
-let timingMode = lsRead(LS_TIMING, "ms");
+let timingMode = "ms";   // the RESTING state; a press peeks at fps and release restores it
 let theme      = lsRead(LS_THEME, "dark");
 
 // ---------------------------------------------------------------------------
@@ -534,7 +532,7 @@ async function sendControl(moduleName, controlName, value) {
     }
     // Toggling expert mode changes which controls RENDER (the `advanced` ones), not just a value: so
     // re-render the cards. Structural change, same as an add/remove; the value write above already landed.
-    if (moduleName === "System" && controlName === "expertMode") renderCards();
+    if (moduleName === "System" && controlName === "mode") renderCards();
     // Best-effort by design: failures are not retried here. Non-ok responses +
     // network errors are logged to console so a user with devtools open can see
     // what went wrong (e.g. a control value the device-side validator rejected).
@@ -550,7 +548,7 @@ async function sendControl(moduleName, controlName, value) {
         // The Firmware card's `image` picks WHICH PARTITION every other control describes, so the
         // device rebinds version/build/firmware/partition to the other image and the card has to
         // be redrawn from the new values. Refetched AFTER the POST rather than re-rendered before
-        // it (the expertMode case above), because the new values only exist once the device has
+        // it (the mode case above), because the new values only exist once the device has
         // switched. The routine WS push cannot carry it: renderCards is suppressed while the user
         // is interacting, and operating this select is exactly that.
         else if (moduleName === "Firmware" && controlName === "image") refetchState();
@@ -981,13 +979,15 @@ function findModule(name, modules) {
     return null;
 }
 
-// Global "expert mode": the System module's expertMode control. Controls tagged `advanced` (dev/tuning
-// readouts + knobs) render only when this is on. Read live from state so a toggle takes effect on the
-// next render with no reload; default off if System or the control isn't present yet.
-function isExpertMode() {
+// How much the reader wants to see: the System module's `mode` control, 0 user, 1 expert, 2
+// developer. A control carries the mode it needs (`minMode`) and renders when the reader is at
+// least there. Read live from state so a change takes effect on the next render with no reload;
+// 0 when System or the control isn't present yet, so a device mid-boot shows the show and nothing
+// that needs explaining.
+function uiMode() {
     const sys = state ? findModule("System") : null;
-    const c = sys && sys.controls && sys.controls.find(c => c.name === "expertMode");
-    return !!(c && c.value);
+    const c = sys && sys.controls && sys.controls.find(c => c.name === "mode");
+    return c ? (c.value | 0) : 0;
 }
 
 function renderCards() {
@@ -1751,37 +1751,45 @@ function createCard(mod, depth) {
     // showing ✓ when on, blank when off. Stores its checked state in
     // data-checked so updateValues can sync from WS pushes. A native <input>
     // would not match the other buttons' frame and corner radius.
-    const enabled = document.createElement("button");
-    enabled.type = "button";
-    enabled.className = "module-enabled";
-    enabled.dataset.mid = mod.name;
-    enabled.dataset.key = "enabled";
-    enabled.setAttribute("aria-pressed", "true");
-    enabled.title = "Enable / disable";
-    const setEnabledUi = (on) => {
-        enabled.dataset.checked = on ? "true" : "false";
-        enabled.textContent = "⏻";
-        enabled.classList.toggle("module-enabled--off", !on);
-        enabled.setAttribute("aria-pressed", on ? "true" : "false");
-        card.classList.toggle("card--disabled", !on);
-        // Grey this module's TAB in the same click, alongside its card: so the tab title dims INSTANTLY
-        // instead of waiting ~1s for the server's full-state round-trip. (updateTabDot still syncs it on the
-        // patch path, idempotently, so this just makes the on/off button the immediate driver.) The tab
-        // lives in the parent's strip, found by the same data-tab-mid updateTabDot uses.
-        const tabEl = queryByName(`.tab[data-tab-mid="${cssEscape(mod.name)}"]`, "data-tab-mid", mod.name);
-        if (tabEl) tabEl.classList.toggle("tab--disabled", !on);
-    };
-    setEnabledUi(mod.enabled === undefined ? true : !!mod.enabled);
-    enabled.addEventListener("click", () => {
-        const next = enabled.dataset.checked !== "true";
-        setEnabledUi(next);
-        // Stamp dragTs so a WS state push older than this click can't revert
-        // the toggle before the server has acknowledged. updateValues reads
-        // dragTs[mod.name + ":enabled"] on line ~952 and suppresses stale
-        // patches within the 1s cooldown.
-        dragTs[mod.name + ":enabled"] = Date.now();
-        sendControl(mod.name, "enabled", next);
-    });
+    // Only where it DOES something. A module that keeps running regardless says so
+    // (`respectsEnabled`), and on Network, System or Firmware the switch was a control that
+    // changed nothing: the scheduler ticks them whatever the flag says, so the device can be
+    // reached to re-enable whatever else was turned off.
+    // Built only where it does something; the name, tags and stats below are every card's.
+    let enabled = null;
+    if (mod.respectsEnabled !== false) {
+        enabled = document.createElement("button");
+        enabled.type = "button";
+        enabled.className = "module-enabled";
+        enabled.dataset.mid = mod.name;
+        enabled.dataset.key = "enabled";
+        enabled.setAttribute("aria-pressed", "true");
+        enabled.title = "Enable / disable";
+        const setEnabledUi = (on) => {
+            enabled.dataset.checked = on ? "true" : "false";
+            enabled.textContent = "⏻";
+            enabled.classList.toggle("module-enabled--off", !on);
+            enabled.setAttribute("aria-pressed", on ? "true" : "false");
+            card.classList.toggle("card--disabled", !on);
+            // Grey this module's TAB in the same click, alongside its card: so the tab title dims INSTANTLY
+            // instead of waiting ~1s for the server's full-state round-trip. (updateTabDot still syncs it on the
+            // patch path, idempotently, so this just makes the on/off button the immediate driver.) The tab
+            // lives in the parent's strip, found by the same data-tab-mid updateTabDot uses.
+            const tabEl = queryByName(`.tab[data-tab-mid="${cssEscape(mod.name)}"]`, "data-tab-mid", mod.name);
+            if (tabEl) tabEl.classList.toggle("tab--disabled", !on);
+        };
+        setEnabledUi(mod.enabled === undefined ? true : !!mod.enabled);
+        enabled.addEventListener("click", () => {
+            const next = enabled.dataset.checked !== "true";
+            setEnabledUi(next);
+            // Stamp dragTs so a WS state push older than this click can't revert
+            // the toggle before the server has acknowledged. updateValues reads
+            // dragTs[mod.name + ":enabled"] on line ~952 and suppresses stale
+            // patches within the 1s cooldown.
+            dragTs[mod.name + ":enabled"] = Date.now();
+            sendControl(mod.name, "enabled", next);
+        });
+    }
 
     const name = document.createElement("span");
     name.className = "card-name";
@@ -1803,27 +1811,38 @@ function createCard(mod, depth) {
     spacer.className = "card-spacer";
     title.appendChild(spacer);
 
-    // fps/ms toggle on the stats line: global mode, single click cycles all cards
-    const stats = document.createElement("span");
-    stats.className = "card-stats";
-    stats.dataset.mid = mod.name;
-    stats.dataset.key = "stats";
-    stats.title = formatStatsTitle(mod);
-    stats.textContent = formatStats(mod);
-    stats.addEventListener("click", () => {
-        const idx = TIMING_MODES.indexOf(timingMode);
-        timingMode = TIMING_MODES[(idx + 1) % TIMING_MODES.length];
-        localStorage.setItem(LS_TIMING, timingMode);
-        // Refresh every card's stats line in place: no full re-render needed
-        document.querySelectorAll(".card-stats[data-mid]").forEach(s => {
-            const m = findModule(s.dataset.mid);
-            if (m) { s.textContent = formatStats(m); s.title = formatStatsTitle(m); }
-        });
-    });
-    title.appendChild(stats);
+    // Built from expert up: in user mode the line is empty and there is nothing to hold.
+    //
+    // HOLD TO PEEK, the password field's gesture. Time is what the line says, because a tick time
+    // is the honest per-module figure: fps inverts it, so a module ticking in 22 us reads as 45K
+    // fps and invites being read as a frame rate it is not. The rate is one press away for anyone
+    // who wants it, and it goes back on release rather than leaving every card in a mode.
+    if (uiMode() >= 1) {
+        const stats = document.createElement("span");
+        stats.className = "card-stats";
+        stats.dataset.mid = mod.name;
+        stats.dataset.key = "stats";
+        stats.title = formatStatsTitle(mod);
+        stats.textContent = formatStats(mod);
+        const peekAll = (mode) => {
+            timingMode = mode;
+            document.querySelectorAll(".card-stats[data-mid]").forEach(el => {
+                const m = findModule(el.dataset.mid);
+                if (m) { el.textContent = formatStats(m); el.title = formatStatsTitle(m); }
+            });
+        };
+        const show = () => peekAll("fps");
+        const hide = () => peekAll("ms");
+        stats.addEventListener("mousedown", show);
+        stats.addEventListener("mouseup", hide);
+        stats.addEventListener("mouseleave", hide);
+        stats.addEventListener("touchstart", (e) => { e.preventDefault(); show(); });
+        stats.addEventListener("touchend", hide);
+        title.appendChild(stats);
+    }
 
     // Enable checkbox joins the right-hand action cluster, before ✎/×.
-    title.appendChild(enabled);
+    if (enabled) title.appendChild(enabled);
 
     // Delete / replace buttons for user-managed children (any role a container
     // accepts, minus modules that opted out via userEditable=false). Top-level
@@ -2317,7 +2336,12 @@ function fmtBytes(n) {
 // Stats line: timing (🕒, fps or µs/ms per the global toggle) + memory
 // (🧠 static, plus "+ dynamic" only when the module allocated heap).
 // Timing is omitted entirely when the module has no measured loop time.
+//
+// EXPERT AND UP. A tick time and a byte count diagnose the firmware rather than the show: they
+// say which module is spending the frame, which is a question a reader in user mode is not
+// asking and cannot act on. The same gate the controls use, so one mode decides the whole card.
 function formatStats(mod) {
+    if (uiMode() < 1) return "";
     const us = (mod.tickTimeUs !== undefined) ? mod.tickTimeUs : 0;
     let timing = "";
     if (us > 0) {
@@ -2528,7 +2552,7 @@ function renderEmojiTags(el, list) {
 // paths (renderCards's initial build + updateModuleControls's WS live-patch) so they agree.
 function controlRendersGenerically(mod, ctrl) {
     if (ctrl.hidden) return false;
-    if (ctrl.advanced && !isExpertMode()) return false;   // expert-only control, expert mode is off
+    if ((ctrl.minMode | 0) > uiMode()) return false;   // above the reader's mode
     return true;
 }
 
@@ -2537,7 +2561,10 @@ function createControl(moduleName, moduleType, ctrl) {
     row.className = "control-row";
     // Expert-only controls (only reachable here when expert mode is on: see controlRendersGenerically)
     // get a distinct treatment so they read as a different tier: a left accent stripe + muted label.
-    if (ctrl.advanced) row.classList.add("control-advanced");
+    // One class per tier, because the glyph says WHICH mode reveals it: a reader in expert mode
+    // meets both marks and the developer one tells them what they are still not seeing.
+    if ((ctrl.minMode | 0) === 1) row.classList.add("control-expert");
+    else if ((ctrl.minMode | 0) >= 2) row.classList.add("control-developer");
     // A switch-row control renders as a strip like an encoder or fader, so switch N sits in the
     // same column as encoder N and fader N: a surface reads down a channel, not across a list.
     if (ctrl.switchRow) row.classList.add("control-switch");
@@ -2545,10 +2572,10 @@ function createControl(moduleName, moduleType, ctrl) {
 
     const label = document.createElement("label");
     label.className = "control-label";
-    // The expertMode toggle itself carries the wrench glyph (via CSS ::before), so it reads as the switch
+    // The mode select itself carries both glyphs (via CSS ::after), so it reads as the switch
     // that governs the 🔧 controls: the toggle is never `advanced` (it must always be reachable), so key
     // it by name rather than the flag.
-    if (moduleName === "System" && ctrl.name === "expertMode") label.classList.add("control-label--expert");
+    if (moduleName === "System" && ctrl.name === "mode") label.classList.add("control-label--mode");
     label.textContent = displayName(ctrl.name);
     // The display strip carries no label: it spans the card and shows whatever was last touched, so
     // a label naming one control would be wrong the moment another moved.
diff --git a/src/ui/style.css b/src/ui/style.css
index a85e561d..7f94f6d9 100644
--- a/src/ui/style.css
+++ b/src/ui/style.css
@@ -725,12 +725,22 @@ body.cards-resizing {
     color: var(--fg-muted);
 }
 
-/* Expert-only controls (ctrl.advanced, shown only in expert mode): a small wrench glyph after the label
-   marks them as a distinct "tuning/dev" tier at a glance — semantically clearer than a bare accent stripe.
-   The same glyph sits on the expertMode toggle itself (see .control-label--expert), tying cause to effect. */
-.control-advanced > .control-label::after,
-.control-label--expert::after {
+/* A control above user mode carries a glyph naming the tier that reveals it, which reads at a glance
+   where an accent stripe would not. The SLIDER is expert, the tuning a working installation needs;
+   the WRENCH is developer, what diagnoses the firmware. The mode select itself carries both, tying
+   cause to effect (see .control-label--mode). */
+.control-expert > .control-label::after {
+    content: "🎚️";
+}
+.control-developer > .control-label::after {
     content: "🔧";
+}
+.control-label--mode::after {
+    content: "🎚️🔧";
+}
+.control-expert > .control-label::after,
+.control-developer > .control-label::after,
+.control-label--mode::after {
     font-size: 10px;
     margin-left: 4px;
     opacity: 0.7;
diff --git a/test/python/test_check_docgen.py b/test/python/test_check_docgen.py
index c0a74b09..2dc66945 100644
--- a/test/python/test_check_docgen.py
+++ b/test/python/test_check_docgen.py
@@ -917,6 +917,60 @@ def test_a_lead_is_measured_as_a_block_so_chopping_does_not_satisfy_it():
         "no single line is over the cap, which is the point of measuring the block"
 
 
+def test_the_ratchet_refuses_a_rule_that_rose_against_the_committed_report(monkeypatch):
+    """A warning is staged work, and staged work that grows is not a sweep. The committed report
+    is the number to beat, per RULE and on the total: per rule alone misses a rule still under
+    its own baseline while the total rises, and the total alone hides one rule paying for
+    another, which is what splitting a long line does to the width and block counts.
+
+    Drives `_ratchet` itself. An earlier version of this test asserted against a local copy of the
+    comparison and passed while the real function compared internal keys against display labels,
+    so it matched nothing and the per-rule half never fired."""
+    import check_docgen
+    base = {"(total)": 10, "multi-line comment blocks": 4, "hard wraps": 2}
+    monkeypatch.setattr(check_docgen, "_committed_counts", lambda: base)
+
+    # A `.cpp` key, because the ratchet guards the WARNING column and a header's finding blocks.
+    cpp = "src/core/x.cpp"
+    block = "code comment 3 lines > 1"
+
+    assert check_docgen._ratchet([(cpp, block)] * 3) == [], "under every baseline: silent"
+
+    rose = check_docgen._ratchet([(cpp, block)] * 5)
+    assert rose == [("multi-line comment blocks", 4, 5)], \
+        f"one rule paying for another is caught, and the total (5) stays under 10: {rose}"
+
+    # The labels the report writes ARE the keys the comparison uses, in both directions.
+    assert check_docgen._rule_label(check_docgen._rule_name(block)) in base
+
+    # And the real parser finds the tracked report's numbers, or the comparison reads nothing.
+    monkeypatch.undo()
+    counts = check_docgen._committed_counts()
+    assert counts is None or "(total)" in counts, counts
+    # Only the `## By rule` table feeds the baseline: an area row would be a rule that never appears.
+    assert counts is None or not any(k.startswith("`") for k in counts), counts
+
+
+def test_a_hard_wrap_blocks_in_an_implementation_file_too():
+    """The one rule staged the OTHER way. Every other `.cpp` finding warns while the sweep runs,
+    but the tree is at ZERO hard wraps, so there is nothing left to stage and the exemption only
+    protects a regression. The reason holds everywhere too: a split sentence reflows every line
+    it spans on the next word change, which costs a reviewer the same in either file."""
+    import check_docgen
+    src = ("#include <x.h>\n\nvoid g();\n\n"
+           "// A sentence that carries on\n"
+           "// onto the following line.\n"
+           "void f();\n")
+    for rel in ("src/core/x.h", "src/core/x.cpp"):
+        hits = [(k, w) for k, w in check_docgen._header_rules(rel, src) if "hard wrap" in w]
+        assert hits, rel
+        assert all(check_docgen._blocks(k, w) for k, w in hits), (rel, "must BLOCK")
+    # And the rules still being swept keep warning in a `.cpp`, or the staging is gone entirely.
+    runs = [(k, w) for k, w in check_docgen._header_rules("src/core/x.cpp", src)
+            if "code comment" in w]
+    assert runs and not any(check_docgen._blocks(k, w) for k, w in runs), runs
+
+
 def test_the_line_length_cap_warns_even_in_a_header():
     """The one staged exception. Every other rule describes how the tree is already written, so
     a finding is a defect against a standard the file was written under. This cap is new, so its
diff --git a/test/scenarios/core/scenario_MoonModule_control_change.json b/test/scenarios/core/scenario_MoonModule_control_change.json
index 05831d52..27c460be 100644
--- a/test/scenarios/core/scenario_MoonModule_control_change.json
+++ b/test/scenarios/core/scenario_MoonModule_control_change.json
@@ -117,14 +117,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 126,
+            "p50": 124,
             "p95": 196,
-            "min": 117,
+            "min": 115,
             "max": 198,
             "n": 32,
-            "samples": [160, 120, 161, 163, 194, 186, 133, 122, 196, 123, 194, 119, 121, 126, 126, 121, 123, 117, 194, 117, 123, 121, 124, 194, 124, 193, 191, 192, 198, 194, 142, 119]
+            "samples": [163, 194, 186, 133, 122, 196, 123, 194, 119, 121, 126, 126, 121, 123, 117, 194, 117, 123, 121, 124, 194, 124, 193, 191, 192, 198, 194, 142, 119, 115, 120, 119]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
@@ -304,9 +304,9 @@
             "min": 119,
             "max": 139,
             "n": 32,
-            "samples": [123, 120, 125, 135, 131, 129, 131, 123, 132, 124, 130, 119, 122, 127, 139, 120, 122, 119, 130, 119, 128, 124, 121, 131, 139, 130, 131, 129, 132, 131, 127, 122]
+            "samples": [135, 131, 129, 131, 123, 132, 124, 130, 119, 122, 127, 139, 120, 122, 119, 130, 119, 128, 124, 121, 131, 139, 130, 131, 129, 132, 131, 127, 122, 120, 120, 119]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
@@ -481,14 +481,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 119,
+            "p50": 120,
             "p95": 125,
             "min": 115,
             "max": 126,
             "n": 32,
-            "samples": [118, 118, 120, 116, 118, 117, 120, 121, 120, 121, 115, 119, 125, 125, 118, 121, 121, 120, 117, 119, 123, 126, 120, 117, 123, 118, 116, 115, 117, 116, 116, 122]
+            "samples": [116, 118, 117, 120, 121, 120, 121, 115, 119, 125, 125, 118, 121, 121, 120, 117, 119, 123, 126, 120, 117, 123, 118, 116, 115, 117, 116, 116, 122, 121, 121, 120]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
@@ -671,14 +671,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 120,
-            "p95": 126,
+            "p50": 119,
+            "p95": 125,
             "min": 117,
             "max": 128,
             "n": 32,
-            "samples": [121, 126, 122, 117, 119, 119, 120, 121, 121, 119, 118, 119, 122, 125, 120, 120, 121, 120, 119, 119, 128, 125, 120, 119, 124, 118, 119, 118, 119, 120, 119, 119]
+            "samples": [117, 119, 119, 120, 121, 121, 119, 118, 119, 122, 125, 120, 120, 121, 120, 119, 119, 128, 125, 120, 119, 124, 118, 119, 118, 119, 120, 119, 119, 119, 120, 120]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Audio_mutation.json b/test/scenarios/light/scenario_Audio_mutation.json
index 96eed9bc..c3e7b447 100644
--- a/test/scenarios/light/scenario_Audio_mutation.json
+++ b/test/scenarios/light/scenario_Audio_mutation.json
@@ -109,9 +109,9 @@
             "min": 16,
             "max": 39,
             "n": 32,
-            "samples": [17, 16, 16, 16, 16, 16, 20, 16, 20, 16, 19, 18, 21, 16, 16, 20, 16, 16, 17, 39, 17, 20, 20, 17, 16, 16, 16, 17, 16, 17, 17, 17]
+            "samples": [16, 16, 16, 20, 16, 20, 16, 19, 18, 21, 16, 16, 20, 16, 16, 17, 39, 17, 20, 20, 17, 16, 16, 16, 17, 16, 17, 17, 17, 20, 20, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -202,13 +202,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 21,
-            "p95": 67,
+            "p95": 43,
             "min": 17,
             "max": 186,
             "n": 32,
-            "samples": [67, 37, 18, 17, 25, 19, 19, 21, 18, 20, 21, 18, 20, 17, 43, 41, 27, 25, 23, 26, 36, 17, 20, 186, 21, 33, 21, 18, 26, 20, 18, 17]
+            "samples": [17, 25, 19, 19, 21, 18, 20, 21, 18, 20, 17, 43, 41, 27, 25, 23, 26, 36, 17, 20, 186, 21, 33, 21, 18, 26, 20, 18, 17, 39, 26, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -316,13 +316,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 19,
-            "p95": 65,
+            "p95": 48,
             "min": 17,
             "max": 109,
             "n": 32,
-            "samples": [65, 29, 20, 18, 28, 18, 20, 17, 19, 19, 21, 19, 20, 17, 19, 34, 24, 19, 20, 30, 30, 18, 23, 109, 29, 32, 19, 17, 19, 19, 18, 17]
+            "samples": [18, 28, 18, 20, 17, 19, 19, 21, 19, 20, 17, 19, 34, 24, 19, 20, 30, 30, 18, 23, 109, 29, 32, 19, 17, 19, 19, 18, 17, 26, 48, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -413,13 +413,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 22,
-            "p95": 47,
+            "p95": 38,
             "min": 18,
             "max": 61,
             "n": 32,
-            "samples": [47, 19, 22, 19, 38, 22, 20, 18, 24, 24, 22, 23, 21, 22, 20, 61, 21, 21, 23, 34, 28, 22, 21, 21, 27, 28, 20, 19, 20, 21, 23, 18]
+            "samples": [19, 38, 22, 20, 18, 24, 24, 22, 23, 21, 22, 20, 61, 21, 21, 23, 34, 28, 22, 21, 21, 27, 28, 20, 19, 20, 21, 23, 18, 26, 29, 23]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -512,9 +512,9 @@
             "min": 18,
             "max": 50,
             "n": 32,
-            "samples": [28, 21, 20, 18, 26, 21, 20, 18, 35, 20, 23, 21, 21, 22, 21, 22, 19, 19, 20, 21, 50, 22, 19, 19, 40, 21, 19, 21, 23, 21, 22, 21]
+            "samples": [18, 26, 21, 20, 18, 35, 20, 23, 21, 21, 22, 21, 22, 19, 19, 20, 21, 50, 22, 19, 19, 40, 21, 19, 21, 23, 21, 22, 21, 21, 19, 23]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -603,13 +603,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 20,
-            "p95": 23,
-            "min": 17,
-            "max": 33,
+            "p95": 22,
+            "min": 16,
+            "max": 23,
             "n": 32,
-            "samples": [33, 17, 17, 20, 18, 20, 20, 17, 21, 21, 19, 21, 20, 20, 20, 19, 17, 17, 18, 21, 23, 20, 22, 18, 19, 18, 17, 20, 19, 17, 20, 20]
+            "samples": [20, 18, 20, 20, 17, 21, 21, 19, 21, 20, 20, 20, 19, 17, 17, 18, 21, 23, 20, 22, 18, 19, 18, 17, 20, 19, 17, 20, 20, 19, 16, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Aurora_fps.json b/test/scenarios/light/scenario_Aurora_fps.json
index d748b386..9901b245 100644
--- a/test/scenarios/light/scenario_Aurora_fps.json
+++ b/test/scenarios/light/scenario_Aurora_fps.json
@@ -85,13 +85,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 862,
-            "p95": 1266,
+            "p95": 1177,
             "min": 819,
             "max": 1744,
             "n": 32,
-            "samples": [859, 1266, 840, 873, 821, 923, 837, 846, 890, 853, 875, 854, 851, 843, 919, 841, 1005, 1177, 852, 880, 907, 862, 839, 1744, 877, 929, 1001, 834, 899, 864, 835, 819]
+            "samples": [873, 821, 923, 837, 846, 890, 853, 875, 854, 851, 843, 919, 841, 1005, 1177, 852, 880, 907, 862, 839, 1744, 877, 929, 1001, 834, 899, 864, 835, 819, 836, 840, 862]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -114,14 +114,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 322,
+            "p50": 321,
             "p95": 478,
             "min": 308,
             "max": 936,
             "n": 32,
-            "samples": [316, 442, 314, 312, 308, 322, 318, 317, 321, 321, 315, 326, 323, 320, 936, 318, 336, 320, 322, 331, 478, 332, 316, 444, 333, 346, 405, 315, 349, 349, 324, 309]
+            "samples": [312, 308, 322, 318, 317, 321, 321, 315, 326, 323, 320, 936, 318, 336, 320, 322, 331, 478, 332, 316, 444, 333, 346, 405, 315, 349, 349, 324, 309, 316, 316, 317]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -149,9 +149,9 @@
             "min": 180,
             "max": 391,
             "n": 32,
-            "samples": [186, 285, 185, 181, 181, 189, 188, 189, 190, 191, 185, 189, 189, 195, 301, 186, 190, 201, 189, 196, 261, 194, 185, 373, 188, 197, 391, 184, 207, 186, 182, 180]
+            "samples": [181, 181, 189, 188, 189, 190, 191, 185, 189, 189, 195, 301, 186, 190, 201, 189, 196, 261, 194, 185, 373, 188, 197, 391, 184, 207, 186, 182, 180, 185, 185, 185]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -174,14 +174,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 552,
+            "p50": 551,
             "p95": 899,
             "min": 529,
             "max": 1141,
             "n": 32,
-            "samples": [541, 642, 538, 532, 535, 559, 551, 546, 549, 549, 541, 607, 567, 719, 899, 547, 558, 552, 674, 584, 1141, 568, 543, 662, 559, 600, 642, 541, 579, 542, 540, 529]
+            "samples": [532, 535, 559, 551, 546, 549, 549, 541, 607, 567, 719, 899, 547, 558, 552, 674, 584, 1141, 568, 543, 662, 559, 600, 642, 541, 579, 542, 540, 529, 546, 544, 546]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -204,14 +204,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 945,
+            "p50": 940,
             "p95": 1265,
             "min": 910,
             "max": 1683,
             "n": 32,
-            "samples": [934, 1063, 1007, 918, 915, 960, 1015, 940, 945, 949, 938, 934, 977, 1001, 1253, 939, 943, 940, 938, 978, 1683, 1000, 939, 1070, 981, 978, 1265, 931, 998, 932, 917, 910]
+            "samples": [918, 915, 960, 1015, 940, 945, 949, 938, 934, 977, 1001, 1253, 939, 943, 940, 938, 978, 1683, 1000, 939, 1070, 981, 978, 1265, 931, 998, 932, 917, 910, 938, 939, 936]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -234,14 +234,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 1460,
-            "p95": 1678,
+            "p50": 1458,
+            "p95": 1665,
             "min": 1407,
-            "max": 1894,
+            "max": 1678,
             "n": 32,
-            "samples": [1431, 1894, 1531, 1417, 1410, 1461, 1454, 1453, 1462, 1460, 1448, 1443, 1470, 1502, 1486, 1465, 1458, 1460, 1443, 1498, 1555, 1527, 1446, 1503, 1517, 1665, 1678, 1440, 1475, 1440, 1414, 1407]
+            "samples": [1417, 1410, 1461, 1454, 1453, 1462, 1460, 1448, 1443, 1470, 1502, 1486, 1465, 1458, 1460, 1443, 1498, 1555, 1527, 1446, 1503, 1517, 1665, 1678, 1440, 1475, 1440, 1414, 1407, 1445, 1450, 1447]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -264,14 +264,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 1526,
+            "p50": 1521,
             "p95": 1947,
             "min": 1453,
             "max": 1951,
             "n": 32,
-            "samples": [1525, 1655, 1573, 1496, 1498, 1899, 1508, 1648, 1506, 1505, 1505, 1493, 1624, 1526, 1521, 1535, 1500, 1503, 1624, 1564, 1590, 1951, 1590, 1517, 1549, 1679, 1544, 1489, 1947, 1550, 1455, 1453]
+            "samples": [1496, 1498, 1899, 1508, 1648, 1506, 1505, 1505, 1493, 1624, 1526, 1521, 1535, 1500, 1503, 1624, 1564, 1590, 1951, 1590, 1517, 1549, 1679, 1544, 1489, 1947, 1550, 1455, 1453, 1522, 1498, 1497]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -295,13 +295,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 1117,
-            "p95": 1443,
+            "p95": 1258,
             "min": 1071,
             "max": 1828,
             "n": 32,
-            "samples": [1087, 1443, 1096, 1078, 1088, 1131, 1119, 1117, 1109, 1117, 1101, 1099, 1828, 1258, 1115, 1121, 1148, 1103, 1223, 1150, 1163, 1142, 1120, 1116, 1150, 1137, 1134, 1108, 1108, 1177, 1077, 1071]
+            "samples": [1078, 1088, 1131, 1119, 1117, 1109, 1117, 1101, 1099, 1828, 1258, 1115, 1121, 1148, 1103, 1223, 1150, 1163, 1142, 1120, 1116, 1150, 1137, 1134, 1108, 1108, 1177, 1077, 1071, 1106, 1103, 1102]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     }
diff --git a/test/scenarios/light/scenario_Driver_mutation.json b/test/scenarios/light/scenario_Driver_mutation.json
index fb1c7052..92dce12d 100644
--- a/test/scenarios/light/scenario_Driver_mutation.json
+++ b/test/scenarios/light/scenario_Driver_mutation.json
@@ -76,14 +76,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 19,
+            "p50": 20,
             "p95": 30,
             "min": 16,
             "max": 46,
             "n": 32,
-            "samples": [20, 21, 19, 21, 20, 16, 20, 20, 46, 23, 20, 19, 16, 20, 20, 17, 17, 17, 17, 19, 19, 19, 19, 23, 21, 30, 17, 20, 17, 17, 20, 16]
+            "samples": [21, 20, 16, 20, 20, 46, 23, 20, 19, 16, 20, 20, 17, 17, 17, 17, 19, 19, 19, 19, 23, 21, 30, 17, 20, 17, 17, 20, 16, 20, 20, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -178,9 +178,9 @@
             "min": 16,
             "max": 38,
             "n": 32,
-            "samples": [20, 20, 20, 20, 20, 16, 20, 19, 38, 19, 20, 21, 16, 20, 21, 17, 19, 16, 19, 20, 20, 20, 20, 21, 21, 23, 20, 20, 17, 18, 20, 17]
+            "samples": [20, 20, 16, 20, 19, 38, 19, 20, 21, 16, 20, 21, 17, 19, 16, 19, 20, 20, 20, 20, 21, 21, 23, 20, 20, 17, 18, 20, 17, 20, 20, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -275,9 +275,9 @@
             "min": 16,
             "max": 22,
             "n": 32,
-            "samples": [20, 20, 19, 20, 20, 18, 20, 19, 22, 17, 19, 19, 19, 20, 20, 16, 17, 17, 20, 20, 18, 19, 18, 17, 18, 16, 20, 20, 19, 21, 20, 20]
+            "samples": [20, 20, 18, 20, 19, 22, 17, 19, 19, 19, 20, 20, 16, 17, 17, 20, 20, 18, 19, 18, 17, 18, 16, 20, 20, 19, 21, 20, 20, 20, 20, 19]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -370,9 +370,9 @@
             "min": 16,
             "max": 29,
             "n": 32,
-            "samples": [20, 20, 20, 21, 20, 20, 20, 19, 19, 29, 20, 20, 19, 20, 21, 16, 17, 17, 20, 20, 16, 19, 19, 18, 21, 17, 20, 20, 21, 19, 20, 20]
+            "samples": [21, 20, 20, 20, 19, 19, 29, 20, 20, 19, 20, 21, 16, 17, 17, 20, 20, 16, 19, 19, 18, 21, 17, 20, 20, 21, 19, 20, 20, 19, 20, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -460,14 +460,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 19,
+            "p50": 20,
             "p95": 22,
             "min": 16,
             "max": 28,
             "n": 32,
-            "samples": [20, 19, 20, 19, 20, 20, 20, 19, 20, 18, 19, 19, 20, 20, 20, 16, 17, 16, 20, 19, 17, 19, 19, 18, 19, 19, 20, 28, 21, 22, 20, 20]
+            "samples": [19, 20, 20, 20, 19, 20, 18, 19, 19, 20, 20, 20, 16, 17, 16, 20, 19, 17, 19, 19, 18, 19, 19, 20, 28, 21, 22, 20, 20, 20, 20, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Effects_composition.json b/test/scenarios/light/scenario_Effects_composition.json
index 631d4ae5..4ec876fb 100644
--- a/test/scenarios/light/scenario_Effects_composition.json
+++ b/test/scenarios/light/scenario_Effects_composition.json
@@ -111,9 +111,9 @@
             "min": 141,
             "max": 656,
             "n": 32,
-            "samples": [145, 145, 143, 145, 144, 144, 144, 145, 147, 143, 158, 656, 144, 144, 144, 143, 148, 150, 144, 142, 147, 146, 152, 143, 149, 160, 145, 142, 144, 145, 142, 141]
+            "samples": [145, 144, 144, 144, 145, 147, 143, 158, 656, 144, 144, 144, 143, 148, 150, 144, 142, 147, 146, 152, 143, 149, 160, 145, 142, 144, 145, 142, 141, 143, 143, 144]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Fields_polar_lut.json b/test/scenarios/light/scenario_Fields_polar_lut.json
index a53caae2..d49f789e 100644
--- a/test/scenarios/light/scenario_Fields_polar_lut.json
+++ b/test/scenarios/light/scenario_Fields_polar_lut.json
@@ -85,14 +85,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 286,
+            "p50": 285,
             "p95": 369,
-            "min": 274,
+            "min": 275,
             "max": 421,
             "n": 32,
-            "samples": [274, 297, 281, 279, 278, 290, 293, 285, 285, 283, 286, 284, 421, 369, 286, 303, 285, 284, 289, 295, 307, 292, 287, 305, 299, 287, 335, 282, 284, 282, 276, 275]
+            "samples": [279, 278, 290, 293, 285, 285, 283, 286, 284, 421, 369, 286, 303, 285, 284, 289, 295, 307, 292, 287, 305, 299, 287, 335, 282, 284, 282, 276, 275, 282, 283, 283]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -120,9 +120,9 @@
             "min": 274,
             "max": 596,
             "n": 32,
-            "samples": [274, 292, 279, 276, 276, 290, 290, 283, 284, 287, 283, 284, 308, 524, 284, 306, 281, 596, 303, 296, 299, 291, 289, 318, 300, 282, 284, 281, 283, 281, 274, 294]
+            "samples": [276, 276, 290, 290, 283, 284, 287, 283, 284, 308, 524, 284, 306, 281, 596, 303, 296, 299, 291, 289, 318, 300, 282, 284, 281, 283, 281, 274, 294, 282, 285, 284]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -147,12 +147,12 @@
           "tick_us": {
             "p50": 284,
             "p95": 455,
-            "min": 275,
+            "min": 277,
             "max": 732,
             "n": 32,
-            "samples": [275, 337, 278, 277, 281, 286, 283, 284, 283, 289, 280, 282, 297, 372, 732, 284, 282, 313, 295, 376, 319, 292, 291, 455, 317, 279, 307, 281, 282, 281, 278, 301]
+            "samples": [277, 281, 286, 283, 284, 283, 289, 280, 282, 297, 372, 732, 284, 282, 313, 295, 376, 319, 292, 291, 455, 317, 279, 307, 281, 282, 281, 278, 301, 285, 283, 283]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -180,9 +180,9 @@
             "min": 273,
             "max": 607,
             "n": 32,
-            "samples": [274, 334, 281, 275, 278, 283, 282, 286, 287, 283, 282, 281, 297, 324, 343, 284, 303, 307, 304, 371, 305, 607, 289, 326, 302, 273, 282, 282, 281, 282, 276, 274]
+            "samples": [275, 278, 283, 282, 286, 287, 283, 282, 281, 297, 324, 343, 284, 303, 307, 304, 371, 305, 607, 289, 326, 302, 273, 282, 282, 281, 282, 276, 274, 283, 283, 282]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -205,14 +205,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 141,
-            "p95": 253,
-            "min": 135,
+            "p50": 140,
+            "p95": 250,
+            "min": 136,
             "max": 408,
             "n": 32,
-            "samples": [135, 253, 140, 139, 136, 141, 139, 140, 141, 141, 139, 139, 146, 156, 142, 140, 149, 197, 147, 149, 159, 408, 250, 141, 145, 143, 140, 140, 140, 139, 136, 136]
+            "samples": [139, 136, 141, 139, 140, 141, 141, 139, 139, 146, 156, 142, 140, 149, 197, 147, 149, 159, 408, 250, 141, 145, 143, 140, 140, 140, 139, 136, 136, 140, 140, 139]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -235,14 +235,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 434,
-            "p95": 622,
-            "min": 407,
+            "p50": 424,
+            "p95": 562,
+            "min": 408,
             "max": 641,
             "n": 32,
-            "samples": [407, 622, 421, 411, 414, 439, 421, 420, 446, 463, 419, 418, 434, 492, 424, 424, 470, 641, 447, 442, 562, 485, 488, 435, 449, 439, 420, 422, 415, 439, 408, 411]
+            "samples": [411, 414, 439, 421, 420, 446, 463, 419, 418, 434, 492, 424, 424, 470, 641, 447, 442, 562, 485, 488, 435, 449, 439, 420, 422, 415, 439, 408, 411, 421, 421, 421]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -271,14 +271,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 1275,
-            "p95": 2014,
+            "p50": 1263,
+            "p95": 1624,
             "min": 1198,
-            "max": 2429,
+            "max": 2014,
             "n": 32,
-            "samples": [1237, 2429, 1286, 1229, 1219, 1531, 1219, 1451, 1439, 1624, 1275, 1237, 1551, 1340, 1223, 1253, 2014, 1265, 1319, 1459, 1282, 1421, 1365, 1251, 1287, 1262, 1209, 1309, 1263, 1253, 1198, 1218]
+            "samples": [1229, 1219, 1531, 1219, 1451, 1439, 1624, 1275, 1237, 1551, 1340, 1223, 1253, 2014, 1265, 1319, 1459, 1282, 1421, 1365, 1251, 1287, 1262, 1209, 1309, 1263, 1253, 1198, 1218, 1251, 1239, 1224]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -301,14 +301,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 477,
+            "p50": 476,
             "p95": 698,
             "min": 457,
             "max": 1029,
             "n": 32,
-            "samples": [471, 540, 479, 464, 461, 478, 473, 474, 475, 479, 476, 482, 482, 520, 474, 479, 698, 509, 483, 501, 1029, 473, 532, 472, 475, 488, 468, 525, 473, 471, 477, 457]
+            "samples": [464, 461, 478, 473, 474, 475, 479, 476, 482, 482, 520, 474, 479, 698, 509, 483, 501, 1029, 473, 532, 472, 475, 488, 468, 525, 473, 471, 477, 457, 473, 471, 491]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     }
diff --git a/test/scenarios/light/scenario_Fluid_solver.json b/test/scenarios/light/scenario_Fluid_solver.json
index 331267de..10826896 100644
--- a/test/scenarios/light/scenario_Fluid_solver.json
+++ b/test/scenarios/light/scenario_Fluid_solver.json
@@ -83,9 +83,9 @@
             "min": 28,
             "max": 76,
             "n": 32,
-            "samples": [28, 33, 29, 28, 28, 29, 29, 29, 30, 29, 30, 29, 33, 31, 29, 29, 76, 29, 29, 30, 30, 30, 35, 30, 30, 29, 29, 33, 29, 29, 32, 28]
+            "samples": [28, 28, 29, 29, 29, 30, 29, 30, 29, 33, 31, 29, 29, 76, 29, 29, 30, 30, 30, 35, 30, 30, 29, 29, 33, 29, 29, 32, 28, 29, 29, 30]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -105,9 +105,9 @@
             "min": 18,
             "max": 196,
             "n": 32,
-            "samples": [19, 24, 19, 18, 19, 19, 19, 19, 19, 19, 19, 19, 21, 21, 19, 19, 196, 19, 19, 20, 23, 19, 24, 19, 19, 19, 19, 21, 19, 19, 32, 18]
+            "samples": [18, 19, 19, 19, 19, 19, 19, 19, 19, 21, 21, 19, 19, 196, 19, 19, 20, 23, 19, 24, 19, 19, 19, 19, 21, 19, 19, 32, 18, 19, 19, 19]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -127,9 +127,9 @@
             "min": 64,
             "max": 472,
             "n": 32,
-            "samples": [66, 73, 66, 67, 66, 66, 66, 66, 66, 67, 66, 67, 72, 72, 67, 67, 115, 66, 66, 75, 71, 74, 77, 67, 70, 66, 66, 82, 67, 67, 472, 64]
+            "samples": [67, 66, 66, 66, 66, 66, 67, 66, 67, 72, 72, 67, 67, 115, 66, 66, 75, 71, 74, 77, 67, 70, 66, 66, 82, 67, 67, 472, 64, 66, 66, 66]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -149,9 +149,9 @@
             "min": 28,
             "max": 87,
             "n": 32,
-            "samples": [29, 31, 28, 29, 28, 29, 29, 29, 29, 29, 29, 29, 32, 32, 29, 28, 87, 30, 29, 30, 30, 29, 38, 29, 30, 29, 29, 38, 29, 28, 40, 28]
+            "samples": [29, 28, 29, 29, 29, 29, 29, 29, 29, 32, 32, 29, 28, 87, 30, 29, 30, 30, 29, 38, 29, 30, 29, 29, 38, 29, 28, 40, 28, 29, 29, 29]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -171,9 +171,9 @@
             "min": 61,
             "max": 118,
             "n": 32,
-            "samples": [62, 69, 63, 61, 61, 64, 62, 62, 63, 63, 62, 64, 65, 70, 62, 63, 118, 63, 62, 67, 66, 67, 72, 63, 69, 62, 63, 80, 62, 62, 74, 62]
+            "samples": [61, 61, 64, 62, 62, 63, 63, 62, 64, 65, 70, 62, 63, 118, 63, 62, 67, 66, 67, 72, 63, 69, 62, 63, 80, 62, 62, 74, 62, 64, 63, 63]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -192,9 +192,9 @@
             "min": 125,
             "max": 175,
             "n": 32,
-            "samples": [127, 164, 128, 126, 127, 129, 128, 130, 129, 129, 127, 166, 131, 146, 127, 129, 160, 127, 127, 134, 134, 134, 175, 128, 135, 127, 129, 159, 128, 127, 153, 125]
+            "samples": [126, 127, 129, 128, 130, 129, 129, 127, 166, 131, 146, 127, 129, 160, 127, 127, 134, 134, 134, 175, 128, 135, 127, 129, 159, 128, 127, 153, 125, 127, 128, 127]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       },
       "description": "And the height, making it 64x64: four times the cells of the 32x32 the pair started from. Reallocating on each axis separately is the shape a UI resize actually takes."
@@ -211,13 +211,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 129,
-            "p95": 157,
+            "p95": 155,
             "min": 127,
             "max": 157,
             "n": 32,
-            "samples": [128, 157, 128, 129, 127, 129, 129, 129, 129, 130, 128, 130, 133, 157, 129, 129, 149, 129, 129, 136, 134, 129, 136, 129, 134, 128, 128, 141, 128, 128, 155, 127]
+            "samples": [129, 127, 129, 129, 129, 129, 130, 128, 130, 133, 157, 129, 129, 149, 129, 129, 136, 134, 129, 136, 129, 134, 128, 128, 141, 128, 128, 155, 127, 128, 129, 128]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -233,13 +233,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 129,
-            "p95": 223,
+            "p95": 184,
             "min": 125,
             "max": 472,
             "n": 32,
-            "samples": [127, 223, 130, 125, 128, 129, 127, 130, 129, 472, 127, 130, 131, 143, 129, 128, 184, 128, 127, 133, 136, 128, 136, 131, 136, 128, 127, 138, 132, 127, 177, 127]
+            "samples": [125, 128, 129, 127, 130, 129, 472, 127, 130, 131, 143, 129, 128, 184, 128, 127, 133, 136, 128, 136, 131, 136, 128, 127, 138, 132, 127, 177, 127, 127, 128, 129]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -259,9 +259,9 @@
             "min": 34,
             "max": 47,
             "n": 32,
-            "samples": [34, 39, 35, 35, 35, 35, 35, 35, 35, 34, 34, 37, 35, 37, 34, 35, 47, 35, 34, 36, 36, 35, 37, 35, 38, 34, 35, 38, 39, 34, 37, 34]
+            "samples": [35, 35, 35, 35, 35, 35, 34, 34, 37, 35, 37, 34, 35, 47, 35, 34, 36, 36, 35, 37, 35, 38, 34, 35, 38, 39, 34, 37, 34, 34, 35, 35]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -280,9 +280,9 @@
             "min": 10,
             "max": 26,
             "n": 32,
-            "samples": [10, 11, 11, 10, 10, 10, 10, 10, 10, 11, 10, 11, 11, 11, 11, 10, 26, 10, 10, 11, 11, 11, 11, 10, 11, 10, 10, 14, 10, 10, 11, 10]
+            "samples": [10, 10, 10, 10, 10, 10, 11, 10, 11, 11, 11, 11, 10, 26, 10, 10, 11, 11, 11, 11, 10, 11, 10, 10, 14, 10, 10, 11, 10, 10, 10, 10]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -301,9 +301,9 @@
             "min": 214,
             "max": 335,
             "n": 32,
-            "samples": [214, 227, 216, 214, 218, 217, 220, 216, 216, 217, 215, 231, 221, 235, 215, 245, 335, 219, 214, 224, 231, 218, 231, 218, 225, 216, 216, 250, 215, 214, 216, 214]
+            "samples": [214, 218, 217, 220, 216, 216, 217, 215, 231, 221, 235, 215, 245, 335, 219, 214, 224, 231, 218, 231, 218, 225, 216, 216, 250, 215, 214, 216, 214, 216, 216, 216]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -323,9 +323,9 @@
             "min": 10,
             "max": 12,
             "n": 32,
-            "samples": [11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 11, 11, 10, 11, 11, 11, 11, 12, 10, 11, 10, 10, 11, 10, 11, 10, 11]
+            "samples": [11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 11, 11, 10, 11, 11, 11, 11, 12, 10, 11, 10, 10, 11, 10, 11, 10, 11, 11, 11, 10]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -345,9 +345,9 @@
             "min": 8,
             "max": 9,
             "n": 32,
-            "samples": [8, 9, 9, 9, 8, 8, 9, 8, 8, 8, 8, 9, 9, 9, 8, 9, 9, 8, 8, 9, 9, 8, 9, 8, 8, 8, 8, 9, 8, 8, 9, 8]
+            "samples": [9, 8, 8, 9, 8, 8, 8, 8, 9, 9, 9, 8, 9, 9, 8, 8, 9, 9, 8, 9, 8, 8, 8, 8, 9, 8, 8, 9, 8, 8, 8, 9]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -366,9 +366,9 @@
             "min": 6,
             "max": 8,
             "n": 32,
-            "samples": [7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 7, 7, 7, 7, 7, 7, 7, 7, 8, 7, 6, 7, 7]
+            "samples": [7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 7, 7, 7, 7, 7, 7, 7, 7, 8, 7, 6, 7, 7, 7, 6, 7]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     }
diff --git a/test/scenarios/light/scenario_GridBlacks_blackpixel.json b/test/scenarios/light/scenario_GridBlacks_blackpixel.json
index 7828398c..44508e4c 100644
--- a/test/scenarios/light/scenario_GridBlacks_blackpixel.json
+++ b/test/scenarios/light/scenario_GridBlacks_blackpixel.json
@@ -97,7 +97,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -200,7 +200,7 @@
             "n": 32,
             "samples": [2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_GridLayout_resize.json b/test/scenarios/light/scenario_GridLayout_resize.json
index 0add933b..ce0df001 100644
--- a/test/scenarios/light/scenario_GridLayout_resize.json
+++ b/test/scenarios/light/scenario_GridLayout_resize.json
@@ -122,9 +122,9 @@
             "min": 117,
             "max": 158,
             "n": 32,
-            "samples": [117, 126, 127, 117, 120, 121, 120, 118, 123, 124, 120, 124, 126, 132, 120, 142, 128, 119, 118, 124, 118, 119, 129, 121, 126, 117, 119, 158, 122, 120, 117, 118]
+            "samples": [117, 120, 121, 120, 118, 123, 124, 120, 124, 126, 132, 120, 142, 128, 119, 118, 124, 118, 119, 129, 121, 126, 117, 119, 158, 122, 120, 117, 118, 120, 119, 119]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
@@ -304,9 +304,9 @@
             "min": 59,
             "max": 86,
             "n": 32,
-            "samples": [63, 64, 70, 64, 68, 65, 61, 65, 65, 66, 64, 68, 66, 64, 64, 76, 86, 74, 65, 60, 60, 64, 66, 64, 61, 64, 65, 65, 64, 64, 59, 64]
+            "samples": [64, 68, 65, 61, 65, 65, 66, 64, 68, 66, 64, 64, 76, 86, 74, 65, 60, 60, 64, 66, 64, 61, 64, 65, 65, 64, 64, 59, 64, 64, 64, 65]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
@@ -486,9 +486,9 @@
             "min": 116,
             "max": 263,
             "n": 32,
-            "samples": [119, 126, 126, 118, 120, 120, 124, 119, 122, 121, 120, 125, 121, 120, 119, 128, 263, 119, 119, 126, 118, 120, 126, 122, 124, 120, 120, 123, 119, 120, 116, 119]
+            "samples": [118, 120, 120, 124, 119, 122, 121, 120, 125, 121, 120, 119, 128, 263, 119, 119, 126, 118, 120, 126, 122, 124, 120, 120, 123, 119, 120, 116, 119, 120, 119, 119]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth-wifi": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Layer_base_pipeline.json b/test/scenarios/light/scenario_Layer_base_pipeline.json
index 1d4897f1..6d097ceb 100644
--- a/test/scenarios/light/scenario_Layer_base_pipeline.json
+++ b/test/scenarios/light/scenario_Layer_base_pipeline.json
@@ -88,9 +88,9 @@
             "min": 63,
             "max": 87,
             "n": 32,
-            "samples": [72, 66, 66, 70, 71, 67, 71, 68, 70, 66, 71, 70, 70, 70, 66, 87, 70, 69, 71, 67, 70, 74, 70, 69, 69, 71, 72, 70, 69, 64, 63, 67]
+            "samples": [70, 71, 67, 71, 68, 70, 66, 71, 70, 70, 70, 66, 87, 70, 69, 71, 67, 70, 74, 70, 69, 69, 71, 72, 70, 69, 64, 63, 67, 71, 68, 69]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Layer_memory_1to1.json b/test/scenarios/light/scenario_Layer_memory_1to1.json
index b1cc7d06..25cafdb0 100644
--- a/test/scenarios/light/scenario_Layer_memory_1to1.json
+++ b/test/scenarios/light/scenario_Layer_memory_1to1.json
@@ -85,9 +85,9 @@
             "min": 4,
             "max": 28,
             "n": 32,
-            "samples": [5, 7, 5, 4, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 24, 8, 5, 5, 5, 28, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5]
+            "samples": [4, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 24, 8, 5, 5, 5, 28, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Layouts_mutation.json b/test/scenarios/light/scenario_Layouts_mutation.json
index 0a1dd627..cff36acb 100644
--- a/test/scenarios/light/scenario_Layouts_mutation.json
+++ b/test/scenarios/light/scenario_Layouts_mutation.json
@@ -83,9 +83,9 @@
             "min": 16,
             "max": 45,
             "n": 32,
-            "samples": [17, 16, 16, 17, 17, 17, 18, 16, 16, 16, 17, 19, 17, 16, 45, 37, 16, 16, 17, 20, 16, 17, 16, 18, 16, 16, 17, 16, 16, 17, 16, 16]
+            "samples": [17, 17, 17, 18, 16, 16, 16, 17, 19, 17, 16, 45, 37, 16, 16, 17, 20, 16, 17, 16, 18, 16, 16, 17, 16, 16, 17, 16, 16, 16, 16, 16]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -206,14 +206,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 50,
+            "p50": 49,
             "p95": 69,
             "min": 44,
             "max": 77,
             "n": 32,
-            "samples": [53, 46, 50, 50, 50, 58, 51, 49, 47, 48, 49, 53, 56, 49, 69, 52, 49, 47, 51, 77, 50, 49, 50, 50, 47, 50, 51, 49, 50, 44, 49, 46]
+            "samples": [50, 50, 58, 51, 49, 47, 48, 49, 53, 56, 49, 69, 52, 49, 47, 51, 77, 50, 49, 50, 50, 47, 50, 51, 49, 50, 44, 49, 46, 49, 46, 49]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -334,9 +334,9 @@
             "min": 90,
             "max": 104,
             "n": 32,
-            "samples": [94, 94, 93, 94, 94, 92, 97, 91, 93, 92, 93, 95, 99, 93, 104, 99, 92, 92, 90, 90, 93, 100, 93, 93, 92, 93, 96, 93, 93, 91, 92, 93]
+            "samples": [94, 94, 92, 97, 91, 93, 92, 93, 95, 99, 93, 104, 99, 92, 92, 90, 90, 93, 100, 93, 93, 92, 93, 96, 93, 93, 91, 92, 93, 92, 93, 94]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -456,9 +456,9 @@
             "min": 16,
             "max": 23,
             "n": 32,
-            "samples": [21, 20, 19, 20, 19, 19, 21, 20, 20, 19, 20, 20, 17, 20, 17, 20, 19, 20, 23, 16, 19, 17, 20, 16, 21, 20, 20, 20, 20, 17, 16, 18]
+            "samples": [20, 19, 19, 21, 20, 20, 19, 20, 20, 17, 20, 17, 20, 19, 20, 23, 16, 19, 17, 20, 16, 21, 20, 20, 20, 20, 17, 16, 18, 19, 20, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_MoonLiveEffect_livescript.json b/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
index f292f8d2..028ff537 100644
--- a/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
+++ b/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
@@ -93,9 +93,9 @@
             "min": 5,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 6, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -211,9 +211,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 4, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 4, 5, 6, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -321,9 +321,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 4, 5, 5, 5, 5, 5, 4, 5, 6, 6, 6, 5, 5, 5, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 4, 5, 6, 6, 6, 5, 5, 5, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -428,12 +428,12 @@
           "tick_us": {
             "p50": 5,
             "p95": 6,
-            "min": 4,
+            "min": 5,
             "max": 7,
             "n": 32,
-            "samples": [5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 5, 5, 6, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 5, 5, 6, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -534,9 +534,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 4, 4, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -637,9 +637,9 @@
             "min": 4,
             "max": 10,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 6, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 4, 6, 7, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 6, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 4, 6, 7, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -740,9 +740,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 4, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 4, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -762,9 +762,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -784,9 +784,9 @@
             "min": 4,
             "max": 8,
             "n": 32,
-            "samples": [6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 6, 5, 5, 6, 5, 5, 5, 8, 5]
+            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 6, 5, 5, 6, 5, 5, 5, 8, 5, 5, 5, 6]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -804,9 +804,9 @@
             "min": 5,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -907,9 +907,9 @@
             "min": 4,
             "max": 9,
             "n": 32,
-            "samples": [6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 4, 7, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 4, 5, 4, 7, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1007,12 +1007,12 @@
           "tick_us": {
             "p50": 5,
             "p95": 6,
-            "min": 5,
+            "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_MoonLive_pipeline.json b/test/scenarios/light/scenario_MoonLive_pipeline.json
index 020c175c..ebf57196 100644
--- a/test/scenarios/light/scenario_MoonLive_pipeline.json
+++ b/test/scenarios/light/scenario_MoonLive_pipeline.json
@@ -379,9 +379,9 @@
             "min": 4,
             "max": 9,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 4, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -536,9 +536,9 @@
             "min": 4,
             "max": 11,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 11, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 11, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -687,9 +687,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 4, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -985,13 +985,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 5,
-            "p95": 9,
+            "p95": 6,
             "min": 5,
             "max": 10,
             "n": 32,
-            "samples": [5, 9, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1127,13 +1127,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 5,
-            "p95": 5,
+            "p95": 6,
             "min": 4,
             "max": 7,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 4, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 4, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json b/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
index 0a6df222..f1687cdc 100644
--- a/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
+++ b/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
@@ -90,13 +90,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 3,
-            "p95": 4,
+            "p95": 3,
             "min": 2,
             "max": 4,
             "n": 32,
-            "samples": [3, 4, 3, 3, 3, 4, 3, 3, 3, 3, 3, 2, 3, 3, 3, 3, 3, 2, 2, 3, 3, 3, 3, 2, 3, 3, 3, 3, 2, 3, 3, 2]
+            "samples": [3, 3, 4, 3, 3, 3, 3, 3, 2, 3, 3, 3, 3, 3, 2, 2, 3, 3, 3, 3, 2, 3, 3, 3, 3, 2, 3, 3, 2, 3, 3, 3]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_MultiplyModifier_pipeline.json b/test/scenarios/light/scenario_MultiplyModifier_pipeline.json
index beea7251..8f75331b 100644
--- a/test/scenarios/light/scenario_MultiplyModifier_pipeline.json
+++ b/test/scenarios/light/scenario_MultiplyModifier_pipeline.json
@@ -94,9 +94,9 @@
             "min": 117,
             "max": 158,
             "n": 32,
-            "samples": [118, 123, 117, 118, 117, 120, 119, 121, 119, 119, 118, 118, 119, 121, 118, 158, 131, 117, 117, 120, 123, 121, 142, 118, 127, 118, 118, 122, 122, 117, 117, 117]
+            "samples": [118, 117, 120, 119, 121, 119, 119, 118, 118, 119, 121, 118, 158, 131, 117, 117, 120, 123, 121, 142, 118, 127, 118, 118, 122, 122, 117, 117, 117, 119, 117, 118]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_Trails_ladder.json b/test/scenarios/light/scenario_Trails_ladder.json
index 9647ba52..671f9f20 100644
--- a/test/scenarios/light/scenario_Trails_ladder.json
+++ b/test/scenarios/light/scenario_Trails_ladder.json
@@ -91,9 +91,9 @@
             "min": 11,
             "max": 13,
             "n": 32,
-            "samples": [11, 11, 11, 11, 11, 12, 11, 11, 11, 11, 11, 11, 11, 13, 12, 12, 11, 11, 11, 11, 13, 11, 11, 11, 12, 11, 11, 12, 11, 11, 11, 11]
+            "samples": [11, 11, 12, 11, 11, 11, 11, 11, 11, 11, 13, 12, 12, 11, 11, 11, 11, 13, 11, 11, 11, 12, 11, 11, 12, 11, 11, 11, 11, 11, 11, 11]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -128,9 +128,9 @@
             "min": 44,
             "max": 55,
             "n": 32,
-            "samples": [45, 45, 45, 44, 44, 46, 45, 44, 46, 47, 44, 45, 46, 47, 47, 46, 47, 45, 44, 46, 47, 44, 46, 44, 46, 45, 45, 55, 46, 44, 45, 44]
+            "samples": [44, 44, 46, 45, 44, 46, 47, 44, 45, 46, 47, 47, 46, 47, 45, 44, 46, 47, 44, 46, 44, 46, 45, 45, 55, 46, 44, 45, 44, 44, 45, 45]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -165,9 +165,9 @@
             "min": 171,
             "max": 274,
             "n": 32,
-            "samples": [177, 182, 180, 177, 175, 178, 177, 179, 178, 178, 176, 177, 189, 189, 187, 180, 184, 176, 176, 185, 191, 178, 189, 178, 191, 176, 176, 274, 188, 178, 198, 171]
+            "samples": [177, 175, 178, 177, 179, 178, 178, 176, 177, 189, 189, 187, 180, 184, 176, 176, 185, 191, 178, 189, 178, 191, 176, 176, 274, 188, 178, 198, 171, 176, 176, 178]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -204,14 +204,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 348,
+            "p50": 347,
             "p95": 421,
             "min": 335,
             "max": 466,
             "n": 32,
-            "samples": [345, 355, 348, 345, 346, 356, 347, 347, 348, 347, 466, 347, 351, 358, 348, 399, 362, 348, 346, 358, 365, 346, 361, 345, 359, 346, 347, 421, 355, 346, 362, 335]
+            "samples": [345, 346, 356, 347, 347, 348, 347, 466, 347, 351, 358, 348, 399, 362, 348, 346, 358, 365, 346, 361, 345, 359, 346, 347, 421, 355, 346, 362, 335, 347, 347, 347]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -234,14 +234,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 359,
+            "p50": 358,
             "p95": 658,
             "min": 348,
             "max": 742,
             "n": 32,
-            "samples": [351, 367, 359, 358, 352, 359, 353, 354, 356, 356, 352, 352, 359, 523, 376, 365, 368, 658, 362, 367, 371, 354, 370, 354, 369, 355, 358, 488, 359, 354, 742, 348]
+            "samples": [358, 352, 359, 353, 354, 356, 356, 352, 352, 359, 523, 376, 365, 368, 658, 362, 367, 371, 354, 370, 354, 369, 355, 358, 488, 359, 354, 742, 348, 355, 355, 357]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -269,9 +269,9 @@
             "min": 345,
             "max": 426,
             "n": 32,
-            "samples": [345, 359, 347, 352, 347, 350, 346, 348, 400, 349, 346, 346, 351, 426, 347, 357, 349, 350, 354, 364, 360, 346, 358, 346, 370, 351, 349, 393, 352, 345, 349, 345]
+            "samples": [352, 347, 350, 346, 348, 400, 349, 346, 346, 351, 426, 347, 357, 349, 350, 354, 364, 360, 346, 358, 346, 370, 351, 349, 393, 352, 345, 349, 345, 347, 347, 360]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     },
@@ -309,13 +309,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 177,
-            "p95": 187,
+            "p95": 186,
             "min": 176,
             "max": 268,
             "n": 32,
-            "samples": [176, 187, 177, 178, 177, 177, 177, 179, 268, 178, 176, 176, 179, 186, 176, 178, 177, 182, 176, 184, 179, 177, 183, 177, 185, 176, 177, 177, 178, 176, 176, 176]
+            "samples": [178, 177, 177, 177, 179, 268, 178, 176, 176, 179, 186, 176, 178, 177, 182, 176, 184, 179, 177, 183, 177, 185, 176, 177, 177, 178, 176, 176, 176, 177, 176, 184]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         }
       }
     }
diff --git a/test/scenarios/light/scenario_modifier_chain.json b/test/scenarios/light/scenario_modifier_chain.json
index 5b67b99f..4fd587ad 100644
--- a/test/scenarios/light/scenario_modifier_chain.json
+++ b/test/scenarios/light/scenario_modifier_chain.json
@@ -107,9 +107,9 @@
             "min": 8,
             "max": 27,
             "n": 32,
-            "samples": [8, 10, 10, 8, 8, 10, 10, 9, 10, 8, 10, 9, 11, 9, 10, 10, 10, 9, 10, 8, 10, 27, 9, 10, 10, 8, 10, 10, 10, 8, 8, 10]
+            "samples": [8, 8, 10, 10, 9, 10, 8, 10, 9, 11, 9, 10, 10, 10, 9, 10, 8, 10, 27, 9, 10, 10, 8, 10, 10, 10, 8, 8, 10, 10, 10, 10]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -167,9 +167,9 @@
             "min": 6,
             "max": 11,
             "n": 32,
-            "samples": [7, 9, 8, 9, 7, 9, 7, 9, 9, 6, 9, 9, 9, 9, 9, 9, 9, 8, 9, 7, 9, 11, 9, 9, 9, 9, 8, 9, 10, 7, 9, 8]
+            "samples": [9, 7, 9, 7, 9, 9, 6, 9, 9, 9, 9, 9, 9, 9, 8, 9, 7, 9, 11, 9, 9, 9, 9, 8, 9, 10, 7, 9, 8, 9, 9, 10]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -225,9 +225,9 @@
             "min": 21,
             "max": 25,
             "n": 32,
-            "samples": [24, 25, 22, 24, 22, 24, 21, 25, 25, 21, 24, 25, 24, 24, 24, 24, 24, 24, 24, 21, 24, 24, 25, 25, 25, 24, 24, 24, 25, 21, 23, 24]
+            "samples": [24, 22, 24, 21, 25, 25, 21, 24, 25, 24, 24, 24, 24, 24, 24, 24, 21, 24, 24, 25, 25, 25, 24, 24, 24, 25, 21, 23, 24, 24, 24, 24]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -253,14 +253,14 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 44,
+            "p50": 43,
             "p95": 46,
             "min": 35,
             "max": 47,
             "n": 32,
-            "samples": [46, 44, 45, 44, 38, 44, 47, 42, 43, 45, 43, 45, 44, 43, 43, 43, 42, 44, 45, 37, 44, 44, 41, 43, 43, 44, 39, 45, 45, 36, 35, 46]
+            "samples": [44, 38, 44, 47, 42, 43, 45, 43, 45, 44, 43, 43, 43, 42, 44, 45, 37, 44, 44, 41, 43, 43, 44, 39, 45, 45, 36, 35, 46, 44, 42, 44]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_modifier_swap.json b/test/scenarios/light/scenario_modifier_swap.json
index 089f1770..3e0e42c4 100644
--- a/test/scenarios/light/scenario_modifier_swap.json
+++ b/test/scenarios/light/scenario_modifier_swap.json
@@ -156,9 +156,9 @@
             "min": 8,
             "max": 11,
             "n": 32,
-            "samples": [8, 8, 9, 8, 9, 8, 8, 11, 8, 8, 8, 10, 8, 8, 10, 8, 8, 8, 10, 8, 8, 8, 8, 8, 8, 8, 8, 10, 11, 8, 8, 9]
+            "samples": [8, 9, 8, 8, 11, 8, 8, 8, 10, 8, 8, 10, 8, 8, 8, 10, 8, 8, 8, 8, 8, 8, 8, 8, 10, 11, 8, 8, 9, 8, 8, 8]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth": {
           "tick_us": {
@@ -300,9 +300,9 @@
             "min": 20,
             "max": 53,
             "n": 32,
-            "samples": [21, 23, 23, 23, 24, 21, 21, 24, 21, 21, 24, 24, 25, 24, 24, 24, 21, 23, 25, 20, 23, 22, 22, 23, 24, 21, 53, 24, 25, 21, 21, 24]
+            "samples": [23, 24, 21, 21, 24, 21, 21, 24, 24, 25, 24, 24, 24, 21, 23, 25, 20, 23, 22, 22, 23, 24, 21, 53, 24, 25, 21, 21, 24, 24, 22, 20]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth": {
           "tick_us": {
@@ -444,9 +444,9 @@
             "min": 8,
             "max": 16,
             "n": 32,
-            "samples": [11, 10, 10, 10, 10, 10, 8, 8, 10, 10, 10, 10, 11, 10, 10, 10, 8, 10, 10, 9, 10, 11, 10, 11, 10, 10, 16, 10, 10, 8, 11, 10]
+            "samples": [10, 10, 10, 8, 8, 10, 10, 10, 10, 11, 10, 10, 10, 8, 10, 10, 9, 10, 11, 10, 11, 10, 10, 16, 10, 10, 8, 11, 10, 10, 10, 10]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32-eth": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_perf_full.json b/test/scenarios/light/scenario_perf_full.json
index 0892d06a..d17aa7ee 100644
--- a/test/scenarios/light/scenario_perf_full.json
+++ b/test/scenarios/light/scenario_perf_full.json
@@ -90,9 +90,9 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -210,9 +210,9 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -330,9 +330,9 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -573,9 +573,9 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -693,7 +693,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -822,7 +822,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -955,7 +955,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1062,7 +1062,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
@@ -1175,7 +1175,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1297,9 +1297,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1421,9 +1421,9 @@
             "min": 17,
             "max": 79,
             "n": 32,
-            "samples": [17, 18, 17, 18, 18, 79, 17, 18, 18, 19, 18, 18, 19, 18, 17, 18, 19, 17, 18, 17, 18, 18, 19, 19, 17, 18, 18, 18, 18, 18, 17, 17]
+            "samples": [18, 18, 79, 17, 18, 18, 19, 18, 18, 19, 18, 17, 18, 19, 17, 18, 17, 18, 18, 19, 19, 17, 18, 18, 18, 18, 18, 17, 17, 17, 17, 18]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1545,9 +1545,9 @@
             "min": 70,
             "max": 80,
             "n": 32,
-            "samples": [70, 71, 71, 71, 70, 80, 70, 72, 72, 77, 71, 72, 72, 75, 70, 72, 73, 70, 71, 70, 75, 71, 78, 76, 70, 73, 71, 78, 76, 73, 70, 72]
+            "samples": [71, 70, 80, 70, 72, 72, 77, 71, 72, 72, 75, 70, 72, 73, 70, 71, 70, 75, 71, 78, 76, 70, 73, 71, 78, 76, 73, 70, 72, 70, 70, 70]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1677,9 +1677,9 @@
             "min": 4,
             "max": 5,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1801,9 +1801,9 @@
             "min": 15,
             "max": 18,
             "n": 32,
-            "samples": [15, 16, 15, 16, 15, 16, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 16, 15, 17, 15, 18, 16, 16, 16, 16, 17, 15, 16, 15, 16]
+            "samples": [16, 15, 16, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 16, 15, 17, 15, 18, 16, 16, 16, 16, 17, 15, 16, 15, 16, 16, 15, 16]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -1925,9 +1925,9 @@
             "min": 60,
             "max": 68,
             "n": 32,
-            "samples": [61, 62, 62, 63, 63, 63, 62, 62, 67, 67, 64, 62, 65, 67, 62, 64, 64, 63, 63, 62, 67, 62, 68, 67, 64, 65, 63, 68, 63, 65, 61, 60]
+            "samples": [63, 63, 63, 62, 62, 67, 67, 64, 62, 65, 67, 62, 64, 64, 63, 63, 62, 67, 62, 68, 67, 64, 65, 63, 68, 63, 65, 61, 60, 63, 62, 62]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -2049,9 +2049,9 @@
             "min": 243,
             "max": 297,
             "n": 32,
-            "samples": [243, 252, 250, 252, 251, 251, 250, 252, 265, 297, 250, 250, 266, 295, 249, 257, 261, 251, 251, 251, 265, 251, 269, 272, 252, 262, 251, 263, 248, 265, 251, 243]
+            "samples": [252, 251, 251, 250, 252, 265, 297, 250, 250, 266, 295, 249, 257, 261, 251, 251, 251, 265, 251, 269, 272, 252, 262, 251, 263, 248, 265, 251, 243, 250, 250, 249]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -2210,7 +2210,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32": {
           "tick_us": {
@@ -2334,7 +2334,7 @@
             "n": 32,
             "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32": {
           "tick_us": {
@@ -2456,9 +2456,9 @@
             "min": 15,
             "max": 22,
             "n": 32,
-            "samples": [15, 16, 15, 16, 17, 15, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 15, 15, 16, 15, 16, 22, 17, 17, 15, 15, 15, 16, 17, 15]
+            "samples": [16, 17, 15, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 15, 15, 16, 15, 16, 22, 17, 17, 15, 15, 15, 16, 17, 15, 15, 15, 15]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32": {
           "tick_us": {
@@ -2580,9 +2580,9 @@
             "min": 60,
             "max": 82,
             "n": 32,
-            "samples": [61, 62, 62, 63, 62, 62, 62, 63, 68, 71, 62, 62, 82, 75, 62, 64, 66, 62, 63, 62, 65, 64, 63, 70, 63, 65, 62, 62, 62, 65, 64, 60]
+            "samples": [63, 62, 62, 62, 63, 68, 71, 62, 62, 82, 75, 62, 64, 66, 62, 63, 62, 65, 64, 63, 70, 63, 65, 62, 62, 62, 65, 64, 60, 62, 62, 63]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_perf_light.json b/test/scenarios/light/scenario_perf_light.json
index 452c9bd2..b43518cc 100644
--- a/test/scenarios/light/scenario_perf_light.json
+++ b/test/scenarios/light/scenario_perf_light.json
@@ -108,7 +108,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -456,7 +456,7 @@
             "n": 32,
             "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -578,9 +578,9 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
@@ -702,9 +702,9 @@
             "min": 15,
             "max": 27,
             "n": 32,
-            "samples": [16, 15, 15, 15, 15, 15, 15, 15, 15, 16, 17, 27, 15, 15, 16, 18, 15, 16, 16, 15, 16, 15, 16, 17, 16, 16, 15, 15, 21, 17, 16, 15]
+            "samples": [15, 15, 15, 15, 15, 15, 16, 17, 27, 15, 15, 16, 18, 15, 16, 16, 15, 16, 15, 16, 17, 16, 16, 15, 15, 21, 17, 16, 15, 15, 15, 15]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32s3-n16r8": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_peripheral_grid_sweep.json b/test/scenarios/light/scenario_peripheral_grid_sweep.json
index 26b6a621..20766163 100644
--- a/test/scenarios/light/scenario_peripheral_grid_sweep.json
+++ b/test/scenarios/light/scenario_peripheral_grid_sweep.json
@@ -178,9 +178,9 @@
             "min": 4,
             "max": 15,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -305,9 +305,9 @@
             "min": 15,
             "max": 122,
             "n": 32,
-            "samples": [16, 16, 16, 15, 15, 16, 122, 17, 15, 16, 15, 18, 16, 16, 16, 16, 16, 16, 16, 17, 16, 19, 17, 16, 17, 16, 16, 16, 16, 16, 16, 15]
+            "samples": [15, 15, 16, 122, 17, 15, 16, 15, 18, 16, 16, 16, 16, 16, 16, 16, 17, 16, 19, 17, 16, 17, 16, 16, 16, 16, 16, 16, 15, 16, 16, 16]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -432,9 +432,9 @@
             "min": 61,
             "max": 72,
             "n": 32,
-            "samples": [62, 62, 63, 63, 62, 64, 71, 69, 63, 62, 64, 67, 62, 64, 67, 63, 63, 63, 64, 70, 68, 72, 69, 66, 67, 63, 65, 63, 62, 69, 65, 61]
+            "samples": [63, 62, 64, 71, 69, 63, 62, 64, 67, 62, 64, 67, 63, 63, 63, 64, 70, 68, 72, 69, 66, 67, 63, 65, 63, 62, 69, 65, 61, 62, 69, 63]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -559,9 +559,9 @@
             "min": 242,
             "max": 325,
             "n": 32,
-            "samples": [249, 251, 251, 252, 250, 249, 286, 278, 249, 250, 252, 253, 249, 256, 268, 250, 254, 250, 263, 252, 273, 291, 275, 325, 317, 249, 260, 248, 249, 260, 260, 242]
+            "samples": [252, 250, 249, 286, 278, 249, 250, 252, 253, 249, 256, 268, 250, 254, 250, 263, 252, 273, 291, 275, 325, 317, 249, 260, 248, 249, 260, 260, 242, 251, 251, 250]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -707,9 +707,9 @@
             "min": 4,
             "max": 15,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -834,9 +834,9 @@
             "min": 15,
             "max": 19,
             "n": 32,
-            "samples": [15, 15, 15, 15, 15, 17, 17, 17, 15, 15, 16, 15, 15, 16, 16, 15, 16, 15, 16, 16, 16, 18, 17, 19, 16, 16, 15, 15, 15, 17, 16, 15]
+            "samples": [15, 15, 17, 17, 17, 15, 15, 16, 15, 15, 16, 16, 15, 16, 15, 16, 16, 16, 18, 17, 19, 16, 16, 15, 15, 15, 17, 16, 15, 15, 16, 15]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -961,9 +961,9 @@
             "min": 60,
             "max": 81,
             "n": 32,
-            "samples": [63, 63, 63, 62, 62, 62, 67, 67, 62, 62, 62, 63, 62, 64, 67, 63, 63, 62, 68, 64, 66, 73, 67, 81, 68, 62, 62, 61, 63, 65, 63, 60]
+            "samples": [62, 62, 62, 67, 67, 62, 62, 62, 63, 62, 64, 67, 63, 63, 62, 68, 64, 66, 73, 67, 81, 68, 62, 62, 61, 63, 65, 63, 60, 62, 63, 62]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1088,9 +1088,9 @@
             "min": 243,
             "max": 286,
             "n": 32,
-            "samples": [256, 252, 252, 252, 249, 249, 268, 267, 252, 251, 252, 251, 248, 259, 269, 251, 252, 250, 263, 250, 277, 285, 270, 286, 264, 249, 253, 249, 253, 255, 250, 243]
+            "samples": [252, 249, 249, 268, 267, 252, 251, 252, 251, 248, 259, 269, 251, 252, 250, 263, 250, 277, 285, 270, 286, 264, 249, 253, 249, 253, 255, 250, 243, 250, 251, 250]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1238,7 +1238,7 @@
             "n": 32,
             "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1363,9 +1363,9 @@
             "min": 15,
             "max": 17,
             "n": 32,
-            "samples": [16, 15, 15, 16, 15, 16, 17, 16, 15, 15, 16, 15, 16, 16, 16, 15, 15, 15, 16, 15, 17, 17, 16, 17, 16, 16, 16, 15, 15, 16, 15, 15]
+            "samples": [16, 15, 16, 17, 16, 15, 15, 16, 15, 16, 16, 16, 15, 15, 15, 16, 15, 17, 17, 16, 17, 16, 16, 16, 15, 15, 16, 15, 15, 15, 15, 15]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1490,9 +1490,9 @@
             "min": 60,
             "max": 70,
             "n": 32,
-            "samples": [62, 62, 63, 62, 64, 62, 66, 65, 63, 62, 63, 62, 62, 63, 66, 62, 63, 62, 67, 62, 69, 68, 65, 70, 65, 62, 62, 61, 62, 64, 61, 60]
+            "samples": [62, 64, 62, 66, 65, 63, 62, 63, 62, 62, 63, 66, 62, 63, 62, 67, 62, 69, 68, 65, 70, 65, 62, 62, 61, 62, 64, 61, 60, 62, 65, 62]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1612,14 +1612,14 @@
         },
         "desktop-macos": {
           "tick_us": {
-            "p50": 252,
+            "p50": 251,
             "p95": 278,
             "min": 245,
             "max": 395,
             "n": 32,
-            "samples": [251, 252, 253, 251, 249, 249, 254, 263, 251, 250, 251, 267, 253, 261, 265, 250, 254, 250, 265, 252, 275, 278, 255, 258, 395, 250, 248, 246, 250, 256, 245, 251]
+            "samples": [251, 249, 249, 254, 263, 251, 250, 251, 267, 253, 261, 265, 250, 254, 250, 265, 252, 275, 278, 255, 258, 395, 250, 248, 246, 250, 256, 245, 251, 250, 250, 249]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1765,9 +1765,9 @@
             "min": 4,
             "max": 10,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 10]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 10, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -1887,14 +1887,14 @@
         },
         "desktop-macos": {
           "tick_us": {
-            "p50": 16,
+            "p50": 15,
             "p95": 17,
             "min": 15,
             "max": 32,
             "n": 32,
-            "samples": [16, 15, 15, 15, 15, 15, 16, 16, 15, 15, 15, 17, 15, 17, 16, 16, 15, 16, 16, 15, 17, 17, 16, 16, 17, 15, 15, 16, 15, 16, 15, 32]
+            "samples": [15, 15, 15, 16, 16, 15, 15, 15, 17, 15, 17, 16, 16, 15, 16, 16, 15, 17, 17, 16, 16, 17, 15, 15, 16, 15, 16, 15, 32, 15, 15, 15]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -2014,14 +2014,14 @@
         },
         "desktop-macos": {
           "tick_us": {
-            "p50": 62,
-            "p95": 72,
+            "p50": 63,
+            "p95": 77,
             "min": 60,
             "max": 114,
             "n": 32,
-            "samples": [62, 62, 62, 62, 62, 62, 63, 68, 62, 64, 62, 68, 62, 65, 65, 62, 63, 62, 64, 62, 69, 67, 63, 66, 72, 61, 62, 63, 62, 64, 60, 114]
+            "samples": [62, 62, 62, 63, 68, 62, 64, 62, 68, 62, 65, 65, 62, 63, 62, 64, 62, 69, 67, 63, 66, 72, 61, 62, 63, 62, 64, 60, 114, 62, 62, 77]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
@@ -2142,13 +2142,13 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 254,
-            "p95": 309,
+            "p95": 324,
             "min": 243,
             "max": 349,
             "n": 32,
-            "samples": [249, 251, 258, 252, 247, 250, 264, 260, 251, 349, 250, 268, 248, 257, 266, 250, 254, 248, 262, 246, 273, 264, 254, 263, 284, 258, 248, 254, 249, 260, 243, 309]
+            "samples": [252, 247, 250, 264, 260, 251, 349, 250, 268, 248, 257, 266, 250, 254, 248, 262, 246, 273, 264, 254, 263, 284, 258, 248, 254, 249, 260, 243, 309, 250, 249, 324]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "desktop-windows": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_peripheral_switch.json b/test/scenarios/light/scenario_peripheral_switch.json
index a3780c5d..416a5aa7 100644
--- a/test/scenarios/light/scenario_peripheral_switch.json
+++ b/test/scenarios/light/scenario_peripheral_switch.json
@@ -179,7 +179,7 @@
             "n": 32,
             "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
@@ -300,7 +300,7 @@
             "n": 32,
             "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
@@ -421,7 +421,7 @@
             "n": 32,
             "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
@@ -541,7 +541,7 @@
             "n": 32,
             "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
@@ -660,9 +660,9 @@
             "min": 4,
             "max": 5,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
@@ -797,9 +797,9 @@
             "min": 4,
             "max": 5,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
-          "last_updated": "2026-09-20"
+          "last_updated": "2026-09-21"
         },
         "esp32p4rev1-eth": {
           "tick_us": {
diff --git a/test/unit/light/unit_Drivers_livePalettes.cpp b/test/unit/light/unit_Drivers_livePalettes.cpp
index c9808b79..bad30d20 100644
--- a/test/unit/light/unit_Drivers_livePalettes.cpp
+++ b/test/unit/light/unit_Drivers_livePalettes.cpp
@@ -92,16 +92,12 @@ TEST_CASE("editing a factory palette leaves one entry, not two") {
 }
 
 TEST_CASE("a quote in a live palette's name is escaped, so the options stay parseable") {
-    // A palette is named by its FILE, and a user names the file. An unescaped name ends the JSON
-    // string early and the control frame it sits in stops parsing, so the UI loses the whole
-    // picker rather than one entry. paletteNames() escaped and paletteOptions() did not, for the
-    // same name from the same source.
+    // A palette is named by its FILE, and a user names the file. An unescaped name ends the JSON string early, so the UI loses the whole picker. paletteNames() escaped it and paletteOptions() did not.
     static const char* const kNames[] = {"unit-a\"b"};
     static const char* const kTags[] = {"\U0001F3A8"};
     mm::LivePalettes::set(kNames, kTags, 1);
 
-    // Every built-in with its whole color table, so the dump is tens of kilobytes: a heap buffer
-    // sized past the real thing, so an overflow here means a defect rather than a cap.
+    // Every built-in with its color table is tens of kilobytes, so the buffer is sized past the real thing: an overflow means a defect, not a cap.
     std::vector<char> buf(1 << 17, 0);
     JsonSink sink(buf.data(), buf.size());
     mm::paletteOptions(sink);

From f1be21c42f7b8836874ccfeccf63d229d29b7dd3 Mon Sep 17 00:00:00 2001
From: ewowi <ewowi@icloud.com>
Date: Mon, 21 Sep 2026 12:37:32 +0200
Subject: [PATCH 2/5] Seed the Ethernet preset from the pins after the config
 is restored
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

A board upgrading to the presets now opens on the board it already is, rather than being written back to Custom, and a board that has chosen nothing adopts no preset's pins. Both bugs were found on hardware, on an Olimex, a P4 shield and an S31.

**Core**
- `NetworkModule` seeds `ethBoard` while the selection is still the un-chosen Custom, rather than once on the first `defineControls`. That first render happens BEFORE the saved config is applied, so the seed read the defaults, where `ethType_` is still `ethNone` and no preset carries type 0: it fell through to Custom and the restore persisted that over a board whose pins named a preset.
- `buildEthPresetOptions` falls back to Custom rather than to row 0. Row 0 is a real preset, so a board that had chosen nothing adopted classic ESP32 Ethernet pins, which on an S31 or P4 is the wrong PHY entirely. Invisible to every test and to the desktop preview; found because the test written for the first bug failed with the wrong value.
- The `ethPresetSeeded_` latch goes, replaced by `ethPresetIsUnset()`, which reads the state rather than remembering a visit.

**Scripts/MoonDeck**
- `check_devices` drops an em-dash from two messages.

**Tests**
- A regression test drives the restore sequence the scheduler performs, overlay then rebuild, and pins both bugs. It fails against either of the old behaviours.

**Docs/CI**
- `core/system.md` and `core/ui.md` lose all 98 em-dashes: a definition bullet takes a colon, a mid-sentence aside a comma. Both files were already open for the `mode` control's own entry, and CLAUDE.md names an em-dash as the kind of thing to fix in the edit that meets it.
- `mode` replaces the removed `expertMode` in the System card.

**Notes**
- `previewsEthernetControls` subtracts nothing. It earns its place by making both of these bugs reproducible on the desktop, where the Ethernet controls otherwise compile out and neither could be tested off-hardware.
- Verified on three chip families, each keeping its preset across a reboot with the link up throughout: Olimex ESP32-Gateway on Classic RMII, MHC-WLED P4 shield on P4-NANO, Espressif S31 CoreBoard on S31 CoreBoard. The P4 is the upgrade path itself: it came up from old firmware with saved pins and no `ethBoard`, and recognised them.

**Reviews**
- 🐇 Em-dashes in the `mode` and stats-line bullets → fixed, along with the rest of both files.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 docs/moonmodules/core/system.md               | 130 +++++++++---------
 docs/moonmodules/core/ui.md                   |  72 +++++-----
 moondeck/check/check_devices.py               |   4 +-
 src/core/system/NetworkModule.h               |  20 ++-
 .../unit/core/unit_NetworkModule_ethernet.cpp |  50 +++++++
 5 files changed, 167 insertions(+), 109 deletions(-)

diff --git a/docs/moonmodules/core/system.md b/docs/moonmodules/core/system.md
index c5eb0034..2675647f 100644
--- a/docs/moonmodules/core/system.md
+++ b/docs/moonmodules/core/system.md
@@ -1,6 +1,6 @@
 # Core system
 
-The device's fixed infrastructure — identity, network, provisioning, firmware, and the inspection tools. These modules are **always present and wired by code**, not user-added; the user does not add or delete them. User-added capability modules (Audio, IR) live in the `Services` container instead — see [core/services.md](services.md). Every row links to its generated technical page (the full API, from the `.h`) and its tests. Cross-cutting rationale that no single `.h` owns lives in the prose sections below the table.
+The device's fixed infrastructure: identity, network, provisioning, firmware, and the inspection tools. These modules are **always present and wired by code**, not user-added; the user does not add or delete them. User-added capability modules (Audio, IR) live in the `Services` container instead, see [core/services.md](services.md). Every row links to its generated technical page (the full API, from the `.h`) and its tests. Cross-cutting rationale that no single `.h` owns lives in the prose sections below the table.
 
 ## System modules
 
@@ -8,15 +8,15 @@ The device's fixed infrastructure — identity, network, provisioning, firmware,
 
 ### System
 
-The device's identity and vitals — name (behind mDNS `<name>.local`, the SoftAP SSID, the DHCP hostname), uptime, heap, and per-module footprint reporting. Its fixed inspection children (Tasks, I2C scan) hang beneath it.
+The device's identity and vitals: name (behind mDNS `<name>.local`, the SoftAP SSID, the DHCP hostname), uptime, heap, and per-module footprint reporting. Its fixed inspection children (Tasks, I2C scan) hang beneath it.
 
 <img src="../../assets/core/SystemModule.png" width="300" alt="System module controls">
 
-- `deviceName` — the identity behind mDNS, the SoftAP SSID and the DHCP hostname.
-- `deviceModel` — the board model (drives the installer catalog entry).
-- `mode` — how much of the UI is shown: `user`, `expert` (🎚️) or `developer` (🔧), cumulative.
-- `logLevel` — serial verbosity, defaulting to Warn. The first 60 s always logs at Info.
-- read-only vitals — `uptime`, `fps`, `heap`, `psram`, `flash`, `chip`, and per-module footprint.
+- `deviceName`: the identity behind mDNS, the SoftAP SSID and the DHCP hostname.
+- `deviceModel`: the board model (drives the installer catalog entry).
+- `mode`: how much of the UI is shown, `user`, `expert` (🎚️) or `developer` (🔧), cumulative.
+- `logLevel`: serial verbosity, defaulting to Warn. The first 60 s always logs at Info.
+- read-only vitals: `uptime`, `fps`, `heap`, `psram`, `flash`, `chip`, and per-module footprint.
 
 Detail: [technical](moxygen/SystemModule.md)
 
@@ -30,12 +30,12 @@ WiFi / Ethernet connectivity, static-IP configuration, RSSI and TX-power reporti
 
 <img src="../../assets/core/NetworkModule.png" width="300" alt="Network module controls">
 
-- `mode` — WiFi / Ethernet / off.
-- `ssid` / `password` — WiFi credentials.
-- `mDNS` — the `<name>.local` hostname.
-- `addressing` — DHCP or static; static exposes IP / gateway / subnet / DNS fields.
-- `ethType` / `ethPhyAddr` / `ethRstGpio` / … — Ethernet PHY configuration.
-- read-only — `rssi` (dBm), `txPower` (dBm).
+- `mode`: WiFi / Ethernet / off.
+- `ssid` / `password`: WiFi credentials.
+- `mDNS`: the `<name>.local` hostname.
+- `addressing`: DHCP or static; static exposes IP / gateway / subnet / DNS fields.
+- `ethType` / `ethPhyAddr` / `ethRstGpio` / …, Ethernet PHY configuration.
+- read-only: `rssi` (dBm), `txPower` (dBm).
 
 Detail: [technical](moxygen/NetworkModule.md)
 
@@ -49,7 +49,7 @@ Serial/BLE Improv Wi-Fi provisioning: the web installer hands credentials to a f
 
 <img src="../../assets/core/ImprovProvisioningModule.png" width="300" alt="Improv provisioning module controls">
 
-- `provision_status` — read-only provisioning state.
+- `provision_status`: read-only provisioning state.
 
 Detail: [technical](moxygen/ImprovProvisioningModule.md) · [frame format](moxygen/ImprovFrame.md) · [chunk reassembly](moxygen/ImprovOpReassembler.md)
 
@@ -59,10 +59,10 @@ Detail: [technical](moxygen/ImprovProvisioningModule.md) · [frame format](moxyg
 
 Discovers other projectMM devices on the LAN and lists them, persisting the last-known list across a reboot. A wired-by-code child of Network.
 
-<img src="../../assets/core/DevicesModule.png" width="300" alt="Devices module — discovered LAN devices">
+<img src="../../assets/core/DevicesModule.png" width="300" alt="Devices module, discovered LAN devices">
 
-- `devices` — a List of discovered devices; each row expands to a detail panel. Persistable.
-- `wledCompatible` — also announce on WLED's broadcast address, off by default.
+- `devices`: a List of discovered devices; each row expands to a detail panel. Persistable.
+- `wledCompatible`: also announce on WLED's broadcast address, off by default.
 
 WLED apps browse on broadcast, so a device appears in them only with this on. Off is the better neighbour, since a broadcast wakes every device on the LAN to parse a packet none of them want. Presence always goes to the projectMM group regardless, so peers find each other either way. See [multicast and IGMP snooping](../../explanation/architecture/moonlight.md#multicast-and-igmp-snooping).
 
@@ -78,13 +78,13 @@ Bridges the light to an MQTT broker so a home-automation hub can control it, as
 
 <img src="../../assets/core/MqttModule.png" width="300" alt="MQTT module controls">
 
-- `broker` — the broker hostname (e.g. `homeassistant.lan`) or IP. A hostname is resolved via DNS.
-- `port` — broker port (default 1883).
-- `username` / `password` — broker credentials, optional, the password stored obfuscated.
-- `haDiscovery` — announce a Home Assistant discovery light, off by default.
+- `broker`: the broker hostname (e.g. `homeassistant.lan`) or IP. A hostname is resolved via DNS.
+- `port`: broker port (default 1883).
+- `username` / `password`: broker credentials, optional, the password stored obfuscated.
+- `haDiscovery`: announce a Home Assistant discovery light, off by default.
 
 HA already discovers the device over the WLED shim with no broker, so this stays off to avoid a duplicate entity. Turn it on for broker-only or cross-subnet setups. See the [home-automation guide](../../how-to/home-automation.md).
-- read-only — `mqtt_status`, from `disabled` and `idle` through to `connected`, or an error.
+- read-only: `mqtt_status`, from `disabled` and `idle` through to `connected`, or an error.
 
 Detail: [technical](moxygen/MqttModule.md)
 
@@ -94,13 +94,13 @@ Detail: [technical](moxygen/MqttModule.md)
 
 ### Firmware update
 
-Over-the-air firmware flashing — the one operation that swaps the binary and needs a power cycle (every *config* change applies live; a firmware OTA does not).
+Over-the-air firmware flashing, the one operation that swaps the binary and needs a power cycle (every *config* change applies live; a firmware OTA does not).
 
 <img src="../../assets/core/FirmwareUpdateModule.png" width="300" alt="Firmware update module controls">
 
-- `firmware` — the OTA image to flash.
-- read-only: `version`, `build`, `partition`.
-- `image` — on a device carrying two images, which one those describe and an install writes.
+- `firmware`: the OTA image to flash.
+- read-only: `version`, `build` and `partition`.
+- `image`: on a device carrying two images, which one those describe and an install writes.
 
 The choice is the app it runs, or MoonBase in the factory slot. This control's presence is also what tells the UI that installs run through the reboot-into-MoonBase cycle, behind one "updating firmware" overlay, and that a **Restart in MoonBase** button belongs on the card ([MoonBase](../../explanation/architecture/moonbase.md)).
 
@@ -130,9 +130,9 @@ One opt-in report about this install, sent once per install or upgrade, so devel
 
 <img src="../../assets/core/MoonStatsModule.png" width="300" alt="Stats module controls">
 
-- `consent` — a checkbox, off by default. Nothing is sent and no identifier computed while off.
+- `consent`: a checkbox, off by default. Nothing is sent and no identifier computed while off.
 - read-only: `version` and `reportedVersion`, which differ exactly when a report is due.
-- `send update` — reports again now, for a setup that changed without a version change.
+- `send update`: reports again now, for a setup that changed without a version change.
 
 The two versions differing is what makes an upgrade send one report and a reboot send nothing. The button replaces this install's row rather than adding one.
 
@@ -150,10 +150,10 @@ A public message board between projectMM devices, in the shape Meshtastic's chan
 
 <img src="../../assets/core/MoonTalkModule.png" width="300" alt="Talk module controls">
 
-- `consent` — a checkbox, off by default. Nothing is published or read while it is off.
-- `shareName` — whether your device name rides along, **off by default and a separate decision**.
-- `message` — what to say, up to 280 characters. Typing changes nothing on its own.
-- `send` — publishes the message and clears the box, as does Enter in the message field.
+- `consent`: a checkbox, off by default. Nothing is published or read while it is off.
+- `shareName`: whether your device name rides along, **off by default and a separate decision**.
+- `message`: what to say, up to 280 characters. Typing changes nothing on its own.
+- `send`: publishes the message and clears the box, as does Enter in the message field.
 
 A device name identifies a person rather than a machine. Without it your messages carry the first 8 characters of your installation id, which groups them without naming you.
 
@@ -169,18 +169,18 @@ Detail: [technical](moxygen/MoonTalkModule.md)
 
 Browse and manage the device filesystem: a folder tree with an inline text editor. Distinct from Filesystem, the persistence engine. Behaviour: ⌄ details.
 
-<img src="../../assets/core/FileManagerModule.png" width="300" alt="File Manager panel — folder tree + toolbar">
+<img src="../../assets/core/FileManagerModule.png" width="300" alt="File Manager panel, folder tree + toolbar">
 
-- `file browser` — the panel itself: a folder tree, a toolbar and an inline text editor.
+- `file browser`: the panel itself: a folder tree, a toolbar and an inline text editor.
 - **Backup (⤓)**, download the device's files as one `.json` bundle.
 
 **Keep it private: it contains the WiFi password.** Every file is byte-verified against the listing, and an unreadable one is skipped and named.
 - **Restore (⟲)**, upload a backup bundle, pressing twice since it overwrites the device's files.
 
 Known renames from [MIGRATING.md](../../reference/MIGRATING.md) apply before upload, then a report lists what needs an eye. Every file applies as it lands, bar network settings and the web server's `port`, which the dialog names.
-- `show hidden` — reveal dot-prefixed files and folders, such as `.config`.
-- `filesystem` — read-only usage bar (used / total bytes, from the platform).
-- `lastSaved` — read-only; how long ago config was persisted (read from the Filesystem engine).
+- `show hidden`: reveal dot-prefixed files and folders, such as `.config`.
+- `filesystem`: read-only usage bar (used / total bytes, from the platform).
+- `lastSaved`: read-only; how long ago config was persisted (read from the Filesystem engine).
 
 Detail: [technical](moxygen/FileManagerModule.md)
 
@@ -188,15 +188,15 @@ Detail: [technical](moxygen/FileManagerModule.md)
 
 ### I2C scan
 
-A fixed System module (wired-by-code, always present) that probes the I²C bus on a button press and reports the addresses found — a hardware bring-up tool. The bus pins default to unused (−1), so a board without an I²C device claims no GPIO for it; a board with a bus sets its pins via the catalog, or you type them for an ad-hoc scan. Passive until the scan button is pressed.
+A fixed System module (wired-by-code, always present) that probes the I²C bus on a button press and reports the addresses found, a hardware bring-up tool. The bus pins default to unused (−1), so a board without an I²C device claims no GPIO for it; a board with a bus sets its pins via the catalog, or you type them for an ad-hoc scan. Passive until the scan button is pressed.
 
 <img src="../../assets/core/I2cScanModule.png" width="300" alt="I2C scan module controls">
 
-- `sda` / `scl` — the bus GPIOs, defaulting to −1 for unused.
+- `sda` / `scl`: the bus GPIOs, defaulting to −1 for unused.
 
 A board with a fixed bus injects its own through the catalog, or you type the pins for an ad-hoc scan. The classic Arduino-ESP32 pair is 21/22.
-- `scan` — a button; press to probe the bus now.
-- read-only — `result` (addresses found).
+- `scan`: a button; press to probe the bus now.
+- read-only: `result` (addresses found).
 
 Detail: [technical](moxygen/I2cScanModule.md)
 
@@ -206,10 +206,10 @@ Detail: [technical](moxygen/I2cScanModule.md)
 
 A read-only diagnostic showing **what runs where**: you cannot optimise which module runs on which core until you can see it. A fixed System module, wired-by-code, with each task's MoonModules nested beneath it.
 
-<img src="../../assets/core/TasksModule.png" width="300" alt="Tasks module — a row per FreeRTOS task">
+<img src="../../assets/core/TasksModule.png" width="300" alt="Tasks module, a row per FreeRTOS task">
 
-- read-only — `tasks`, a row per FreeRTOS task.
-- read-only — `core0` / `core1`, what executes on each core, empty on a single-core chip.
+- read-only: `tasks`, a row per FreeRTOS task.
+- read-only: `core0` / `core1`, what executes on each core, empty on a single-core chip.
 
 Each row carries `name`, `state`, `core`, `prio` and `stack`, the minimum free stack seen. A `cpu` percentage appears only in a profiling build, off by default because the run-time counter costs about 5% of the tick.
 
@@ -221,13 +221,13 @@ Detail: [technical](moxygen/TasksModule.md)
 
 ### Pins
 
-A read-only diagnostic showing **which module owns each GPIO, for what role, and whether that pin is safe for it** — the device's pin ownership map, keyed by physical GPIO. A fixed System module, wired-by-code.
+A read-only diagnostic showing **which module owns each GPIO, for what role, and whether that pin is safe for it**, the device's pin ownership map, keyed by physical GPIO. A fixed System module, wired-by-code.
 
 It walks the live tree for every claimed pin, holding no state, and flags double claims.
 
-<img src="../../assets/core/PinsModule.png" width="300" alt="Pins module — the GPIO ownership map">
+<img src="../../assets/core/PinsModule.png" width="300" alt="Pins module, the GPIO ownership map">
 
-- read-only — `pins`, a row per claimed GPIO.
+- read-only: `pins`, a row per claimed GPIO.
 
 Each row carries `gpio`, `owner` and `role`, plus live `dir`, `level` and `drive`. A row takes a coloured edge when unsafe: red for a reserved or double-claimed pin, yellow for a driven role on a strap, per [gpio-usage.md](../../reference/hardware/gpio-usage.md).
 
@@ -237,14 +237,14 @@ Detail: [technical](moxygen/PinsModule.md)
 
 ### Control
 
-A grid of preset pads, a row of rotary encoders above them, a row of on/off switches above those, and a bank of faders below — the layout of a Mackie-style control desk ([X-Touch](https://www.behringer.com/product.html?modelCode=0808-AAF), [QCon Pro G2](https://www.iconproaudio.com/product/qcon-pro-g2/)), so a physical surface maps onto it without a translation layer.
+A grid of preset pads, a row of rotary encoders above them, a row of on/off switches above those, and a bank of faders below, the layout of a Mackie-style control desk ([X-Touch](https://www.behringer.com/product.html?modelCode=0808-AAF), [QCon Pro G2](https://www.iconproaudio.com/product/qcon-pro-g2/)), so a physical surface maps onto it without a translation layer.
 
 <img src="../../assets/core/ControlModule.png" width="300" alt="Control module surface: encoders, preset pads, faders">
 
-- `presets` — one pad per preset file. Click applies, right-click names, drag rearranges.
-- `switch1` … `switch8` — the switch row. `switch1` drives `Drivers.on`, the rest unbound.
-- `encoder1` … `encoder8` — rotary encoders. Drag or scroll to turn, right-click to see the binding.
-- `fader1` … `fader8` — faders. `fader1` drives `Drivers.brightness`, the rest unbound.
+- `presets`: one pad per preset file. Click applies, right-click names, drag rearranges.
+- `switch1` … `switch8`, the switch row. `switch1` drives `Drivers.on`, the rest unbound.
+- `encoder1` … `encoder8`, rotary encoders. Drag or scroll to turn, right-click to see the binding.
+- `fader1` … `fader8`, faders. `fader1` drives `Drivers.brightness`, the rest unbound.
 
 Detail: [technical](moxygen/ControlModule.md)
 
@@ -290,9 +290,9 @@ Detail: [technical](moxygen/ControlSurface.md) · [InputMapping](moxygen/InputMa
 
 #### Presets
 
-A preset is a file: `/.config/presets/<name>.json`. Saving writes one, applying reads one, deleting removes one. Nothing else holds preset state, so there is no second copy to keep in step: the list is rebuilt from the folder rather than persisted alongside it. That rescan runs at startup and after every save, rename and delete — a reorder only rewrites the affected files and re-sorts the rows in place, since the folder's contents have not changed. So a preset added or removed through the File Manager appears once the module next rescans (a reboot, or a save, rename or delete on the surface), not the instant the file lands.
+A preset is a file: `/.config/presets/<name>.json`. Saving writes one, applying reads one, deleting removes one. Nothing else holds preset state, so there is no second copy to keep in step: the list is rebuilt from the folder rather than persisted alongside it. That rescan runs at startup and after every save, rename and delete, a reorder only rewrites the affected files and re-sorts the rows in place, since the folder's contents have not changed. So a preset added or removed through the File Manager appears once the module next rescans (a reboot, or a save, rename or delete on the surface), not the instant the file lands.
 
-The name becomes the file name, so it is restricted to printable ASCII without `/`, `\` or `.` — a validator on the control, which every write path runs. `slot` records which pad the preset occupies, so a surface arranged to match a physical desk survives a reboot.
+The name becomes the file name, so it is restricted to printable ASCII without `/`, `\` or `.`, a validator on the control, which every write path runs. `slot` records which pad the preset occupies, so a surface arranged to match a physical desk survives a reboot.
 
 ##### What a preset carries
 
@@ -320,7 +320,7 @@ A pad is tinted by its role: layout blue, effects violet, driver green, service
 
 ##### Applying is a rebuild
 
-Applying a preset creates, replaces and destroys modules to match what the file describes — it is a restore, not a value overlay: a preset carrying more than the device has adds it, and one describing less removes what it omits.
+Applying a preset creates, replaces and destroys modules to match what the file describes, it is a restore, not a value overlay: a preset carrying more than the device has adds it, and one describing less removes what it omits.
 
 Structural mutation quiesces the render worker, and mutations run inline on the render tick, so a large restore stalls rendering for its duration. The captured subtree is applied and `prepareTree()` runs once at the end. Presets are a cold-path feature; the tick path is untouched.
 
@@ -330,11 +330,11 @@ Looks reach Home Assistant two ways, and only `Effects` presets travel either of
 
 **The WLED integration** (`/presets.json`) is the native path: HA renders looks in its own preset dropdown, shows which one is applied, and applies one when it is chosen. This is what HA calls a preset.
 
-**MQTT discovery** publishes the same looks as the light entity's **effect list**. HA has no preset concept over MQTT, so they arrive as effects — the same result from the user's side, reached through a different mechanism.
+**MQTT discovery** publishes the same looks as the light entity's **effect list**. HA has no preset concept over MQTT, so they arrive as effects, the same result from the user's side, reached through a different mechanism.
 
-HA caches the preset list and re-fetches only when the device's `info.fs.pmt` value changes, so the device reports a revision counter there that bumps on every preset save, rename and delete — a counter rather than a timestamp, so two changes inside one second still read as two. A constant there leaves HA showing the list it read at setup forever; over MQTT the same revision re-announces the effect list mid-session.
+HA caches the preset list and re-fetches only when the device's `info.fs.pmt` value changes, so the device reports a revision counter there that bumps on every preset save, rename and delete, a counter rather than a timestamp, so two changes inside one second still read as two. A constant there leaves HA showing the list it read at setup forever; over MQTT the same revision re-announces the effect list mid-session.
 
-Only looks are exposed, on both paths. A `Drivers` or `Layouts` preset rewires pins or geometry, which must not be reachable from something that believes it is choosing a color scheme — the restriction is enforced at the apply entry point, not merely by omitting them from the list.
+Only looks are exposed, on both paths. A `Drivers` or `Layouts` preset rewires pins or geometry, which must not be reachable from something that believes it is choosing a color scheme, the restriction is enforced at the apply entry point, not merely by omitting them from the list.
 
 Home Assistant's WLED integration connects on **port 80 only**: its host field rejects a port, so a desktop build (which defaults to 8080) needs `--port 80`, and that needs root:
 
@@ -345,7 +345,7 @@ sudo uv run moondeck/run/run_desktop.py --port 80
 The discovery buffers are sized to the looks this device actually has, and grow or shrink as presets are added and removed. There is no cap on the number: a fixed one would either reserve memory a small setup never uses, or silently publish nothing once the list outgrew it.
 
 ## MQTT, details
-The topic prefix is `projectMM/<mac>` — a **stable** identifier (the last 6 hex of the device's MAC), fixed for the device's life. Renaming the device does **not** change its topics, so a hub's config never breaks on a rename (the WLED/Tasmota/Home-Assistant convention). It's derived, not a stored control.
+The topic prefix is `projectMM/<mac>`, a **stable** identifier (the last 6 hex of the device's MAC), fixed for the device's life. Renaming the device does **not** change its topics, so a hub's config never breaks on a rename (the WLED/Tasmota/Home-Assistant convention). It's derived, not a stored control.
 
 **Topics** (for a device whose MAC ends `563cfe`): the device SUBSCRIBEs to the `set` topics and PUBLISHes the `get` topics on change (and on connect, so a controller never reads "No Response"). It also publishes its friendly `deviceName` on the retained `name` topic, so a hub can show the human name while the topics stay MAC-stable:
 
@@ -361,9 +361,9 @@ The topic prefix is `projectMM/<mac>` — a **stable** identifier (the last 6 he
 | device → get | `projectMM/563cfe/update/state` | `{"installed_version":…,"latest_version":…,"release_url":…,"title":…}` (retained; HA update entity) |
 | set → device | `projectMM/563cfe/update/set` | target version string (empty = install latest); triggers OTA against the matching GitHub release asset |
 
-The HomeKit color wheel has no "palette" concept, so `hsv/set`'s hue+saturation pick the **nearest palette** (each built-in palette has a representative color; the closest one is selected) and the value drives brightness — the color wheel becomes a natural palette selector.
+The HomeKit color wheel has no "palette" concept, so `hsv/set`'s hue+saturation pick the **nearest palette** (each built-in palette has a representative color; the closest one is selected) and the value drives brightness, the color wheel becomes a natural palette selector.
 
-**Homebridge** — install [`homebridge-mqttthing`](https://github.com/arachnetech/homebridge-mqttthing) and add a `lightbulb` accessory. Use the device's own MAC suffix (read it from the `mqtt_status`/topics, or `mosquitto_sub -t 'projectMM/#'`) in place of `563cfe`:
+**Homebridge**, install [`homebridge-mqttthing`](https://github.com/arachnetech/homebridge-mqttthing) and add a `lightbulb` accessory. Use the device's own MAC suffix (read it from the `mqtt_status`/topics, or `mosquitto_sub -t 'projectMM/#'`) in place of `563cfe`:
 
 ```json
 {
@@ -387,8 +387,8 @@ The HomeKit color wheel has no "palette" concept, so `hsv/set`'s hue+saturation
 ```
 
 Home Assistant adopts the device two ways, both zero-config:
-- **MQTT auto-discovery** — with `haDiscovery` on (opt-in; off by default) and a broker set, the device announces itself on `homeassistant/light/projectMM_<mac6>/config` and HA auto-creates a wired entity with **on/off + brightness** (the config declares `brightness` only; color isn't in it, so the entity has no color control). Retained across reboots. Color/palette stays on the separate `hsv/set` topic above, not this entity. Off by default because the WLED `/json` shim already gives HA a richer light (color + palette + sensors) over mDNS with no broker — leaving both on lists the device twice; enable this only for broker-only / cross-subnet setups.
-- **WLED integration** — HA's built-in WLED integration discovers the device over the WLED `/json` API projectMM already serves; on/off + brightness work with no broker.
+- **MQTT auto-discovery**: with `haDiscovery` on (opt-in; off by default) and a broker set, the device announces itself on `homeassistant/light/projectMM_<mac6>/config` and HA auto-creates a wired entity with **on/off + brightness** (the config declares `brightness` only; color isn't in it, so the entity has no color control). Retained across reboots. Color/palette stays on the separate `hsv/set` topic above, not this entity. Off by default because the WLED `/json` shim already gives HA a richer light (color + palette + sensors) over mDNS with no broker, leaving both on lists the device twice; enable this only for broker-only / cross-subnet setups.
+- **WLED integration**: HA's built-in WLED integration discovers the device over the WLED `/json` API projectMM already serves; on/off + brightness work with no broker.
 
 Both can be on at once. Setup walkthrough (including exposing HA to Apple Home via HA's HomeKit Bridge, no Homebridge needed) in the [Home Assistant recipe](../../how-to/home-automation.md#adopt-in-home-assistant).
 
@@ -397,8 +397,8 @@ The panel is a lazy folder **tree** (each folder loads its children on first exp
 
 - Click a folder's row to select it and toggle its expansion (▸/▾); click a selected file to open the editor.
 - The toolbar acts on the selected node: **＋ folder** creates a folder inside it, **＋ file** creates an empty file (click it to edit), **🗑 delete** removes the selected file, or a folder and everything inside it (press-twice to confirm), **⟳** refreshes.
-- **Drag files from the desktop** onto a folder (or the tree) to upload them — the body streams straight to the file (any size, binary-safe; capped only by a sanity limit and the free space, which it reports if short); a per-file **⤓** streams it back to the desktop.
+- **Drag files from the desktop** onto a folder (or the tree) to upload them, the body streams straight to the file (any size, binary-safe; capped only by a sanity limit and the free space, which it reports if short); a per-file **⤓** streams it back to the desktop.
 - The editor loads a file's text, pretty-prints JSON on open, and saves atomically; a binary file (contains a NUL) loads read-only (use ⤓ to fetch it intact). Upload and download both stream, so neither truncates.
-- Create / delete are HTTP calls (`POST` / `DELETE /api/dir?path=`), not controls — the path rides the request, so nothing is stored on the device per op.
+- Create / delete are HTTP calls (`POST` / `DELETE /api/dir?path=`), not controls, the path rides the request, so nothing is stored on the device per op.
 
 Last-modified dates (needs an NTP time source + LittleFS mtime), binary/large + folder upload, folder-as-zip download, and `.ml` syntax highlighting are backlogged ([backlog-core § File Manager follow-ups](../../work/future/backlog-core.md#file-manager-follow-ups)).
diff --git a/docs/moonmodules/core/ui.md b/docs/moonmodules/core/ui.md
index 62c0fbf5..79083478 100644
--- a/docs/moonmodules/core/ui.md
+++ b/docs/moonmodules/core/ui.md
@@ -1,9 +1,9 @@
 # Web UI
 
-The projectMM web UI as shipped — the render layer over the MoonModule tree. This page is the UI's
+The projectMM web UI as shipped, the render layer over the MoonModule tree. This page is the UI's
 own implementation spec (status bar, cards, control rendering, styling, the no-rebuild update
-contract). The **high-level architecture** — hand-maintained files, MoonModule-driven rendering, the
-light-domain plug-in points — lives in [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui);
+contract). The **high-level architecture**, hand-maintained files, MoonModule-driven rendering, the
+light-domain plug-in points, lives in [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui);
 the **backend contract** it consumes (every `/api/*` endpoint, the `/ws` frame shape, the control
 descriptors) is owned by [HttpServerModule](moxygen/HttpServerModule.md); the **emoji legend** the
 cards and picker render is [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend).
@@ -36,11 +36,11 @@ This page covers only what those don't: the browser-side rendering behavior.
 └──────────────┴──────────────────────────────────────────────┘
 ```
 
-- Fixed status bar at top (44 px) — hamburger, MoonLight logo, then the rest.
+- Fixed status bar at top (44 px), hamburger, MoonLight logo, then the rest.
 - Side nav: a left column listing root modules; the selected root's card subtree fills the main area
   (one root visible at a time).
 - Module-card column max-width 500 px, centered (single-column, easier to read on phones). The preview
-  canvas is **not** capped — it spans the full content width.
+  canvas is **not** capped, it spans the full content width.
 - Preview canvas is sticky just below the status bar and shrinks 0 % → 50 % as the user scrolls
   0 → 300 px.
 
@@ -48,19 +48,19 @@ This page covers only what those don't: the browser-side rendering behavior.
 
 Fixed top, 44 px, left-to-right:
 
-1. **Hamburger (☰)** — toggles the side nav (`body.nav-open`). See [§ Side navigation](#side-navigation).
-2. **MoonLight logo** — 28 px PNG from `/moonlight-logo.png`; the same image is the page favicon.
-3. **Brand wordmark** — "projectMM".
-4. **Device name** — from the `System.deviceName` control.
-5. **System stats** — `uptime · NN K free` (uptime `Xd Yh Zm Ws`, free heap KB), read from the
+1. **Hamburger (☰)**, toggles the side nav (`body.nav-open`). See [§ Side navigation](#side-navigation).
+2. **MoonLight logo**, 28 px PNG from `/moonlight-logo.png`; the same image is the page favicon.
+3. **Brand wordmark**, "projectMM".
+4. **Device name**, from the `System.deviceName` control.
+5. **System stats**, `uptime · NN K free` (uptime `Xd Yh Zm Ws`, free heap KB), read from the
    SystemModule controls in the WS state push; no separate polling endpoint.
-6. **WebSocket dot** — green = connected, gray = disconnected. The socket auto-reconnects with
+6. **WebSocket dot**, green = connected, gray = disconnected. The socket auto-reconnects with
    exponential backoff (500 ms → 5 s) on close; no manual reconnect button (a page reload covers the
    rare half-dead socket).
-7. **Reboot button (⏻)** — red border via `data-crashed="true"` when `SystemModule.bootReason` is
+7. **Reboot button (⏻)**, red border via `data-crashed="true"` when `SystemModule.bootReason` is
    PANIC / INT_WDT / TASK_WDT / BROWNOUT. Press-twice to confirm: the first click arms it (solid red),
-   a second sends `POST /api/reboot`. Disarms after 3 s or on pointer-leave — no `confirm()` popup.
-8. **Theme toggle (☀/🌙)** — flips `[data-theme]` on `<body>`; preference in `localStorage['mm_theme']`.
+   a second sends `POST /api/reboot`. Disarms after 3 s or on pointer-leave, no `confirm()` popup.
+8. **Theme toggle (☀/🌙)**, flips `[data-theme]` on `<body>`; preference in `localStorage['mm_theme']`.
 
 ## Side navigation
 
@@ -73,32 +73,32 @@ A left column listing the root modules, one entry per top-level MoonModule.
   column the hamburger collapses/expands; on narrow screens (< 820 px) it's a slide-in drawer over a
   dimming overlay (click the overlay or press Esc to close).
 - **The nav states its own order** (`NAV_ORDER` in app.js), which is deliberately NOT the order the
-  roots run in. `main.cpp` orders by dependency — Filesystem before anything that writes a file,
-  System before the modules that read its identity — and that is load-bearing, so it cannot be
+  roots run in. `main.cpp` orders by dependency, Filesystem before anything that writes a file,
+  System before the modules that read its identity, and that is load-bearing, so it cannot be
   reshuffled to suit a menu. It also reads as an implementation detail to a user: it puts System
   first and the lights last. The nav instead groups by what someone is looking for: **Control**,
   then the light pipeline in pipeline order (**Layouts → Effects → Drivers**), then the device
   (**System, File Manager, Network, Services, Firmware**). A root the list does not name still
-  appears, after them, in scheduler order — adding a module never makes it invisible.
+  appears, after them, in scheduler order, adding a module never makes it invisible.
 - **No root reorder.** The order is fixed in that list; the nav does not drag-reorder.
-- **Footer** pinned to the bottom of the nav: social links (GitHub, Discord, Reddit, YouTube — inline
+- **Footer** pinned to the bottom of the nav: social links (GitHub, Discord, Reddit, YouTube, inline
   SVG) and a `© <year> MoonLight` line.
 
 The WS state push carries the full module tree; only the selected root is rendered, and
 `updateValues()` patches just the visible cards (non-rendered roots have no DOM nodes, so their data
-is silently ignored). Per-root server-side filtering was evaluated and deferred — the 1 Hz push is
+is silently ignored). Per-root server-side filtering was evaluated and deferred, the 1 Hz push is
 cheap and the JSON is built through a streaming sink, so tree size is not a buffer-limit concern.
 
 ## Module card
 
-Each MoonModule renders as a card, with **child cards nested inside their parent card's box** — the
+Each MoonModule renders as a card, with **child cards nested inside their parent card's box**, the
 parent's border encloses its children, so the tree shape is visible structurally, not just by
 indentation. Nesting depth shows as progressively lighter backgrounds and a left-margin indent.
 
 ```text
 ┌─ card ──────────────────────────────────┐
 │ [name] [emoji]  [timing · 🧠 mem]  [enabled toggle]  [✎ × ☰]  [? help] [{ } api] │
-│ [control rows — one per control]        │
+│ [control rows, one per control]        │
 │ ┌─ child card ────────────────────────┐ │
 │ │ …                                   │ │
 │ └─────────────────────────────────────┘ │
@@ -109,10 +109,10 @@ indentation. Nesting depth shows as progressively lighter backgrounds and a left
 - The parent's own controls render **above** its children; `+ add child` renders **below** them.
   Child cards live in a `.card-children` wrapper appended into the parent card's DOM node (not flat
   siblings); `renderModuleTree` recurses into the parent card, not into `main`.
-- **`{ }`** opens `GET /api/modules/{name}` in a new tab — that one module's live JSON, for issue
+- **`{ }`** opens `GET /api/modules/{name}` in a new tab, that one module's live JSON, for issue
   reports (see [Log an issue](../../how-to/logging-an-issue.md)). On EVERY card, unlike `✎`/`×` (user-editable
   children only) and `?` (types that have a doc page).
-- **Enabled toggle** in the right-hand action cluster mirrors `MoonModule::enabled()` — a styled
+- **Enabled toggle** in the right-hand action cluster mirrors `MoonModule::enabled()`, a styled
   `<button>` (transparent + muted border, 26×26); state shown by the glyph alone (accent **✓** on,
   muted **○** off; `data-checked` + `aria-pressed`). Toggling fires `onEnabled()`; the Scheduler skips
   disabled modules whose `respectsEnabled()` returns true (default).
@@ -122,7 +122,7 @@ indentation. Nesting depth shows as progressively lighter backgrounds and a left
 - **Help link (?)** at the far right of the title row opens the module's doc page in a new tab. The
   path comes from `docPath` in `/api/types` (engine-provided, relative to `docs/moonmodules/`); omitted
   when the type declares none.
-- **Stats line** — `🕒 <timing>` then `🧠 <static>[ + <dynamic>]`. Timing is a tick time
+- **Stats line**: `🕒 <timing>` then `🧠 <static>[ + <dynamic>]`. Timing is a tick time
   (µs under 1 ms, ms above), omitted when the module has no measured loop time. Memory is
   the C++ object size (`classSize`); the `+ <dynamic>` part (`dynamicBytes`, heap) shows only when the
   module allocated heap. The line shows from expert mode up, and holding it peeks at the rate on every
@@ -131,17 +131,17 @@ indentation. Nesting depth shows as progressively lighter backgrounds and a left
 - **Reorder is drag-and-drop** (HTML5 DnD) on the whole card, desktop and mobile. A drag starting on
   an interactive control is canceled in `dragstart` so the control's own gesture wins; a drag is
   accepted only when source and target share the same `.card-children` container (true siblings).
-- **Replace (✎) / Delete (×) / drag-reorder** apply to *user-managed children* — a module whose role a
+- **Replace (✎) / Delete (×) / drag-reorder** apply to *user-managed children*, a module whose role a
   container accepts (`acceptsChildRoles`) and that hasn't opted out via `userEditable: false`. Replace
   swaps for another type at the same position (picker filtered to the module's role; replacement starts
   at its own defaults). Delete is press-twice confirm. A `userEditable: false` child (e.g.
-  PreviewDriver) shows none of these — it's fixed-shape like a code-wired child.
+  PreviewDriver) shows none of these, it's fixed-shape like a code-wired child.
 - **`+ add child`** renders in a parent's footer when its type declares a non-empty
   `acceptsChildRoles`; it opens the [type picker](#type-picker) filtered to those roles.
 
 ## Control types
 
-Auto-rendered by `controls[].type` — adding a MoonModule with these types needs no UI change. (The
+Auto-rendered by `controls[].type`, adding a MoonModule with these types needs no UI change. (The
 descriptors themselves, and their storage shape, are [Control](moxygen/Control.md) /
 [coding-standards § store values in their native shape](../../contributing/coding-standards.md); this table is
 how each *renders*.)
@@ -183,7 +183,7 @@ how each *renders*.)
   as `{"type":"password","value":"<encoded>"}` where `<encoded>` is the password XOR'd with a fixed key
   then base64-encoded; the UI decodes it (`decodePassword()`) so the input holds the real value (masked,
   hold-to-peek). A **first line of defense only**: the XOR key is a shared constant in both the firmware
-  and `app.js`, so it's trivially reversible — it stops the password being plainly readable in a raw
+  and `app.js`, so it's trivially reversible, it stops the password being plainly readable in a raw
   `curl /api/state`, not a determined reader.
 
 ## Type picker
@@ -192,13 +192,13 @@ One picker serves **add** (`+ add child`) and **replace** (the ✎ button), rend
 (not a modal).
 
 - **Role filter.** Add mode filters to the parent's `acceptsChildRoles` (declared per-type in
-  `/api/types` — the UI hardcodes no container→role map); replace mode filters to the target's own role.
+  `/api/types`, the UI hardcodes no container→role map); replace mode filters to the target's own role.
 - **Emoji tag chips.** A row of toggle chips above the list, one per distinct emoji across the filtered
   types (sources + meanings: [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend)).
-  The UI treats `tags` as opaque — splits it into grapheme clusters, one chip each. Toggling narrows
+  The UI treats `tags` as opaque, splits it into grapheme clusters, one chip each. Toggling narrows
   with **AND** logic (a type shows only if it carries every active chip).
-- **Search box** — substring match on type name; combines with chips (both must match).
-- **Keyboard nav** — type to filter, ↓ to enter list, ↑↓ to move, Enter to confirm, Esc to cancel.
+- **Search box**: substring match on type name; combines with chips (both must match).
+- **Keyboard nav**: type to filter, ↓ to enter list, ↑↓ to move, Enter to confirm, Esc to cancel.
 - **Confirm / Cancel** at the bottom (confirm reads `create` or `replace`); double-click a row to
   confirm immediately.
 
@@ -216,9 +216,9 @@ One picker serves **add** (`+ add child`) and **replace** (the ✎ button), rend
 
 ## Domain preview channel
 
-The UI dedicates a binary slot on its WebSocket — separate from the JSON state updates — for a
+The UI dedicates a binary slot on its WebSocket, separate from the JSON state updates, for a
 domain-specific preview frame; the engine pushes one per render and the UI hands it to a domain
-renderer. Generic shape: `[type-byte] [domain header] [payload]` — the first byte selects the
+renderer. Generic shape: `[type-byte] [domain header] [payload]`, the first byte selects the
 renderer, everything after is the domain's choice; the UI ignores types it doesn't recognise.
 
 - The canvas is sticky just below the status bar, scroll-shrinks 0→50 % over 300 px.
@@ -226,14 +226,14 @@ renderer, everything after is the domain's choice; the UI ignores types it doesn
   caps it.
 - The last received frame is cached so camera gestures (orbit/pan/zoom) redraw instantly without a new
   frame; `touch-action: none` keeps gestures off native scroll/zoom.
-- WebGL clear color `(0, 0, 0, 0)` — a transparent canvas blends into either theme with no per-frame
+- WebGL clear color `(0, 0, 0, 0)`, a transparent canvas blends into either theme with no per-frame
   color work.
 
 The light-domain renderer (WebGL point cloud, the `0x03`/`0x02` frame format, orbit camera,
 downsampling) is [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui) /
 [PreviewDriver](../light/moxygen/PreviewDriver.md).
 
-## State updates — the no-rebuild contract
+## State updates, the no-rebuild contract
 
 When a WS state snapshot arrives, for each module in the payload:
 - If `controls.deviceName` exists, update the status-bar device name and `document.title`.
diff --git a/moondeck/check/check_devices.py b/moondeck/check/check_devices.py
index 0a56d2a5..da8e6af6 100644
--- a/moondeck/check/check_devices.py
+++ b/moondeck/check/check_devices.py
@@ -123,10 +123,10 @@ def main():
     eth_presets = eth_preset_labels()
     for label, field, was, now in eth_preset_drift():
         errors.append(f"NetworkModule.h kEthPresets {label!r}: {field} is {now}, but "
-                      f"platform_config.h's ethConfigDefault says {was} — the preset restates the "
+                      f"platform_config.h's ethConfigDefault says {was}: the preset restates the "
                       f"chip default, so the two must agree or a provisioned board reseeds to Custom")
     if not eth_presets:
-        errors.append("NetworkModule.h: could not read the kEthPresets table — the ethBoard check cannot run")
+        errors.append("NetworkModule.h: could not read the kEthPresets table, so the ethBoard check cannot run")
 
     try:
         catalog = json.loads(CATALOG.read_text(encoding="utf-8"))
diff --git a/src/core/system/NetworkModule.h b/src/core/system/NetworkModule.h
index ba6f91d7..7d91b496 100644
--- a/src/core/system/NetworkModule.h
+++ b/src/core/system/NetworkModule.h
@@ -192,10 +192,10 @@ class NetworkModule : public MoonModule {
             constexpr bool preview = !platform::hasEthernet;
             const uint8_t firstEthControl = controls_.count();
             buildEthPresetOptions();
-            // First render: on a provisioned board the pins are already the catalog's, so read the preset back off them. A factory board has no type yet and lands on Custom with its chip's pins, which is honest until the catalog names a board.
-            if (!ethPresetSeeded_) { ethPresetSeeded_ = true; seedEthPresetFromPins(); }
-            // The restore path fires no onControlChanged, so writing the map HERE is what makes a saved `ethBoard` reach the pins. Keyed on the selection moving, or a rebuild would undo a Custom edit.
-            else if (ethPresetSel_ != ethPresetApplied_) applyEthPreset();
+            // A selection MOVED, so write its map. The restore path fires no onControlChanged, and this is what makes a saved `ethBoard` reach the pins; keyed on the move, or a rebuild would undo a Custom edit.
+            if (ethPresetSel_ != ethPresetApplied_) applyEthPreset();
+            // Otherwise read the preset back off the pins, while the selection is still the un-chosen Custom: restore overlays pins before `ethBoard` survives a rebuild, so seeding once read defaults instead.
+            else if (ethPresetIsUnset()) seedEthPresetFromPins();
             ethPresetApplied_ = ethPresetSel_;
             controls_.addSelect("ethBoard", ethPresetSel_, ethPresetOptions_, ethPresetCount_);
             // By label: the list is filtered per build, so an index would name a different board.
@@ -544,7 +544,6 @@ class NetworkModule : public MoonModule {
     const char* ethPresetOptions_[kEthPresetCount] = {};
     uint8_t ethPresetIndex_[kEthPresetCount] = {};
     uint8_t ethPresetCount_ = 0;
-    bool ethPresetSeeded_ = false;   ///< the preset is matched to the pins once, not every rebuild
     uint8_t ethPresetApplied_ = 0;   ///< the selection whose map is already written, so a rebuild is not a re-apply
 
 
@@ -566,8 +565,11 @@ class NetworkModule : public MoonModule {
             ethPresetIndex_[ethPresetCount_] = i;
             ethPresetCount_++;
         }
-        // By LABEL, so a filtered list cannot silently select a different board.
+        // By LABEL, so a filtered list cannot select a different board, falling back to CUSTOM rather than row 0: a real preset whose map would overwrite the chip's defaults.
         uint8_t sel = 0;
+        for (uint8_t k = 0; k < ethPresetCount_; k++) {
+            if (kEthPresets[ethPresetIndex_[k]].editable) { sel = k; break; }
+        }
         if (current) {
             for (uint8_t k = 0; k < ethPresetCount_; k++) {
                 if (std::strcmp(ethPresetOptions_[k], current) == 0) { sel = k; break; }
@@ -576,6 +578,12 @@ class NetworkModule : public MoonModule {
         ethPresetSel_ = sel;
     }
 
+    /// Has nothing chosen a preset yet? True while the selection is the editable Custom row.
+    bool ethPresetIsUnset() const {
+        if (ethPresetSel_ >= ethPresetCount_) return true;
+        return kEthPresets[ethPresetIndex_[ethPresetSel_]].editable;
+    }
+
     /// Are the pin controls the user's to edit, for the preset currently selected?
     bool ethPinsEditable() const {
         if (ethPresetSel_ >= ethPresetCount_) return true;   // nothing resolved: never hide
diff --git a/test/unit/core/unit_NetworkModule_ethernet.cpp b/test/unit/core/unit_NetworkModule_ethernet.cpp
index 1f765e56..359cf751 100644
--- a/test/unit/core/unit_NetworkModule_ethernet.cpp
+++ b/test/unit/core/unit_NetworkModule_ethernet.cpp
@@ -24,6 +24,7 @@
 #include "platform/platform.h" // setEthConfig / ethStop / ethInit / ethConnected
 #include "core/system/NetworkModule.h"
 #include <cstring>
+#include <string>
 
 // The enum values are a wire contract: the Select index, the ethInit() switch, and every deviceModels.json `ethType` all agree on these. Pin them so a reorder fails here.
 TEST_CASE("EthPhyType enum values match the dropdown/dispatch contract") {
@@ -158,3 +159,52 @@ TEST_CASE("fixedPins never writes past the capacity it is given") {
     CHECK(pads[2].gpio == 0xEE);
     CHECK(net.fixedPins(nullptr, 16) == 0);   // a null sink is answered, not written through
 }
+
+// Two bugs an Olimex found: a board whose pins ARE a preset's map opens on it, and one that chose nothing adopts no preset.
+TEST_CASE("ethBoard seeds from the pins after a restore, not once before it") {
+    if constexpr (!mm::platform::hasEthernet && !mm::platform::previewsEthernetControls) return;
+
+    auto boardOf = [](mm::NetworkModule& n) -> const char* {
+        for (uint8_t i = 0; i < n.controls().count(); i++) {
+            const auto& c = n.controls()[i];
+            if (std::strcmp(c.name, "ethBoard") == 0) {
+                auto* opts = reinterpret_cast<const char* const*>(c.aux);
+                const uint8_t sel = *static_cast<const uint8_t*>(c.ptr);
+                return (opts && sel < c.max) ? opts[sel] : "";
+            }
+        }
+        return nullptr;
+    };
+
+    mm::NetworkModule net;
+    net.rebuildControls();
+    REQUIRE(boardOf(net) != nullptr);
+    // Nothing has chosen one and no type is set, so Custom is the honest answer rather than row 0, a real preset whose map would overwrite the chip's own defaults.
+    CHECK(std::strcmp(boardOf(net), "Custom") == 0);
+
+    // What a restore does, overlaying saved values then rebuilding, with the classic pins and no saved `ethBoard` at all: exactly an upgraded board's config.
+    auto setByName = [&](const char* name, int v) {
+        for (uint8_t i = 0; i < net.controls().count(); i++) {
+            const auto& c = net.controls()[i];
+            if (std::strcmp(c.name, name) != 0) continue;
+            if (c.type == mm::ControlType::Int16) *static_cast<int16_t*>(c.ptr) = static_cast<int16_t>(v);
+            else if (c.type == mm::ControlType::Pin) *static_cast<int8_t*>(c.ptr) = static_cast<int8_t>(v);
+            else *static_cast<uint8_t*>(c.ptr) = static_cast<uint8_t>(v);
+            return;
+        }
+    };
+    setByName("ethType", 1);          // LAN8720
+    setByName("ethPhyAddr", 0);
+    setByName("ethMdcGpio", 23);
+    setByName("ethMdioGpio", 18);
+    setByName("ethRstGpio", 5);
+    setByName("ethClockGpio", 17);
+
+    net.rebuildControls();
+    // The pins name Classic RMII, so the card opens on it rather than on Custom.
+    CHECK(std::strcmp(boardOf(net), "Classic RMII") == 0);
+
+    // And the seed does NOT fight a chosen preset: rebuilding again keeps it.
+    net.rebuildControls();
+    CHECK(std::strcmp(boardOf(net), "Classic RMII") == 0);
+}

From 8f50ae799d5903cf8fad8bc18ac1e6245590e1de Mon Sep 17 00:00:00 2001
From: ewowi <ewowi@icloud.com>
Date: Mon, 21 Sep 2026 21:44:24 +0200
Subject: [PATCH 3/5] Show the whole layout at less detail when the preview
 link is slow
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

A slow link now coarsens the preview instead of shrinking it: the browser keeps every light's position and colors each block from the one light the device sent. Fixes the documentation pipeline dropping seven classes from their pages, and sweeps em-dashes and hard wraps out of the authored docs.

KPI: 16384lights | Desktop:1924KB | tick:124/97/366/2/6/124/276/22/1248/3/259/72/222/17/21/5/1570/124/35/5/144/41/4us(FPS:8064/10309/2732/500000/166666/8064/3623/45454/801/333333/3861/13888/4504/58823/47619/200000/636/8064/28571/200000/6944/24390/250000) | src:270(68891) | test:206(45199) | lizard:267w

**Light domain**
- `PreviewDriver` bounds the coordinate table by MEMORY alone, never by the link. Geometry is static and ships once per epoch at full resolution, while `frameStride()` and `frameCount()` pace the color stream separately. A coarse link therefore costs detail rather than lights.
- The sparse gather filters the cached index by the frame's stride. `keptStep_` records the coarsest lattice each kept light sits on, so a coarser frame filters the cache instead of re-walking `placeLights`, which measured ~8 ms at 12K lights and runs on a nonblocking tick.

**UI**
- `preview3d.js` holds one table per epoch and derives every coarser stride from it, so a stride change costs no table traffic at any rung and the layout keeps its full extent. `upsampleMap()` reproduces the device's kept-light rule, giving each light the color of its block.
- A touch cancelled mid-peek restores the timing display, which `touchend` alone never saw.
- A mode change from another client rebuilds the cards once: it adds or removes rows, which patching values cannot express.
- A control's tier rides a title as well as the CSS glyph, so a screen reader can name what a mark shows.

**Core**
- `presetBuildable` filters Ethernet presets by chip. A classic build offered P4-NANO and S31 CoreBoard, whose pins and PHY it cannot drive.

**Scripts/MoonDeck**
- `gen_api.py` keeps a class the group file never rendered, so seven pages that showed only their free functions carry their class again. `@card` emits the controls screenshot beside a moving preview, and `MM_HEAVY_COMPUTE` is predefined so the two headers behind it generate at all.
- `check_docgen` reports hard wraps in authored markdown as errors, and its ratchet compares the union of baseline and current rules: a rule at zero could rise unseen while the total fell.
- `check_devices` compares the PHY type and clock direction as well as the pins, and reads `ethBoard` by presence so an explicit null is not a preset name.
- `generate_test_docs.py` and the scenario JSONs lose the em-dashes they wrote into the generated test pages.

**Tests**
- A mapped layout at a coarse stride is pinned: the body must carry what the header advertises. It fails against the unfiltered gather.
- Two preview tests move from asserting the table's stride to the frame's, and one pins the split.

**Docs**
- The LED driver card splits into RMT and Parallel, each with its own screenshot and its real controls; six were documented nowhere.
- Twelve control lists corrected against the code, including `targetFps` (1-25, not 1-60), Grid's per-axis ranges, and `curve`, which appeared on no page.
- 750 hard wraps joined across 19 files and the em-dashes removed from the pages this change already touched. The sweep carried content corrections, not only whitespace: the ranges above, and `ui.md`'s description of a state push that the diff model replaced.
- RandomMap is `animated`, not `static`: it keeps a static fold but reshuffles on its own clock.

**Verified on hardware**
- S31 at 16,384 lights: the table ships whole at stride 1 while colors pace at 1024@4, and `Preview` costs 17-19 us per tick.
- A 51,554-light sphere: with the old gather the device shipped 51,554 colors under a header claiming 723; each stride now agrees.

**Reviews**
- 👾 The sparse path shipped the table's count under the frame's header, breaking every mapped layout above stride 1 → fixed, pinned by a test that fails without it.
- 👾 `frameCount()` walked `placeLights` per frame on a nonblocking tick → counted from the cache.
- 👾 The ratchet was violated by this commit's own comments → collapsed, back to baseline.
- 👾 RandomMap animates through `tick()` rather than `modifyLive` → relabelled, and three cards regained the glyph their neighbours carry.
- 🐇 Eight findings: the ratchet's absent-rule hole, preset filtering by chip, the drift check's missing fields, a null `ethBoard`, the mode rebuild, touchcancel, and the tier's accessible name. One was already fixed by the hardware test that found it first.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 .../light/drivers/ParallelLedDriver.png       | Bin 0 -> 34029 bytes
 docs/assets/light/drivers/RmtLedDriver.png    | Bin 13675 -> 27347 bytes
 docs/contributing/principles-and-process.md   |  11 +-
 docs/explanation/architecture/moonbase.md     |  59 +----
 docs/explanation/architecture/mooncore.md     |  14 +-
 docs/friend-repos/FastLED-FastLED.md          |  34 +--
 docs/friend-repos/MoonModules-WLED-MM.md      |  24 +-
 .../PlummersSoftwareLLC-NightDriverStrip.md   |  26 +-
 docs/friend-repos/hpwit-ESPLiveScript.md      |  16 +-
 .../hpwit-I2SClocklessLedDriver.md            |  14 +-
 .../hpwit-I2SClocklessVirtualLedDriver.md     |  18 +-
 docs/friend-repos/hpwit-new-parser.md         |  12 +-
 docs/friend-repos/index.md                    |  55 +++--
 docs/friend-repos/troyhacks-WLED.md           |  22 +-
 docs/friend-repos/wled-WLED.md                |  14 +-
 docs/gettingstarted.md                        | 223 +++++------------
 docs/how-to/building.md                       |  82 ++-----
 docs/how-to/logging-an-issue.md               |  36 +--
 docs/how-to/panel-cards.md                    |  81 ++----
 docs/index.md                                 |   4 +-
 docs/moonmodules/core/services.md             | 100 ++------
 docs/moonmodules/core/system.md               |  14 +-
 docs/moonmodules/core/ui.md                   | 140 +++--------
 docs/moonmodules/light/drivers.md             |  53 ++--
 docs/moonmodules/light/effects.md             | 112 +++++----
 docs/moonmodules/light/layouts.md             |   2 +-
 docs/moonmodules/light/modifiers.md           |  12 +-
 docs/moonmodules/light/moonlive.md            |  81 ++----
 docs/moonmodules/light/power-functions.md     | 101 ++++----
 docs/moonmodules/light/supporting.md          |  38 +--
 docs/reference/MIGRATING.md                   | 168 ++++---------
 docs/reference/hardware/control-surfaces.md   |  25 +-
 .../reference/hardware/esp32-s31-coreboard.md |  61 ++---
 docs/reference/hardware/light-fixtures.md     |  45 +---
 docs/reference/metrics/repo-health.json       |  50 ++--
 docs/reference/metrics/repo-health.md         |  48 ++--
 docs/reference/performance.md                 |  73 ++----
 docs/reference/testing.md                     | 140 +++++------
 docs/tutorials/generative-effects.md          | 230 ++++--------------
 docs/tutorials/how-projectmm-works.md         | 101 +++-----
 moondeck/MoonDeck.md                          |   2 +-
 moondeck/check/check_devices.py               |  18 +-
 moondeck/check/check_docgen.py                |  86 ++++++-
 moondeck/docs/gen_api.py                      |  52 +++-
 moondeck/docs/generate_test_docs.py           |  18 +-
 moondeck/docs/screenshot_modules.py           |   4 +
 src/core/system/MoonCloudModule.h             |  17 +-
 src/core/system/NetworkModule.h               |   9 +-
 src/light/drivers/PreviewDriver.h             |  77 ++++--
 src/light/effects/EchoEffect.h                |   2 +-
 src/ui/app.js                                 |  29 ++-
 src/ui/preview3d.js                           |  80 +++++-
 test/js/ui-live-patch-text.test.mjs           |   9 +-
 test/python/test_check_docgen.py              |  20 ++
 .../scenario_MoonModule_control_change.json   |  10 +-
 ...cenario_MqttModule_haDiscovery_toggle.json |  14 +-
 ...cenario_NetworkModule_eth_reconfigure.json |   8 +-
 .../scenario_NetworkModule_mdns_toggle.json   |   6 +-
 .../light/scenario_Audio_mutation.json        |   6 +-
 .../light/scenario_Driver_mutation.json       |  12 +-
 .../light/scenario_Effects_composition.json   |   8 +-
 .../light/scenario_GridLayout_resize.json     |   2 +-
 .../light/scenario_Layer_base_pipeline.json   |   2 +-
 .../light/scenario_Layouts_mutation.json      |   8 +-
 .../scenario_MoonLiveEffect_controls.json     |  12 +-
 .../scenario_MoonLiveEffect_livescript.json   |  10 +-
 .../light/scenario_MoonLive_pipeline.json     |   2 +-
 .../scenario_MultiplyModifier_memory_lut.json |   2 +-
 .../light/scenario_modifier_chain.json        |   8 +-
 .../light/scenario_modifier_swap.json         |   6 +-
 test/scenarios/light/scenario_perf_full.json  |  16 +-
 test/scenarios/light/scenario_perf_light.json |  12 +-
 .../light/scenario_peripheral_switch.json     |   6 +-
 test/unit/light/unit_PreviewDriver.cpp        |  54 +++-
 74 files changed, 1289 insertions(+), 1677 deletions(-)
 create mode 100644 docs/assets/light/drivers/ParallelLedDriver.png

diff --git a/docs/assets/light/drivers/ParallelLedDriver.png b/docs/assets/light/drivers/ParallelLedDriver.png
new file mode 100644
index 0000000000000000000000000000000000000000..9e179539c00d2da49db8671d85ace65459e959ce
GIT binary patch
literal 34029
zcmce-Wl$V%yRVCDa0@QMArRbU@Zc8Q3GVKLI|PEeOMqa(-QC@t!QF<Sd;aTPtIoS>
z?Q>4m{;)qxQ8m*&-As2s&vjq-?}=1altx1#LV<#ULi;Ksp$Y{B-3oa<Ai+T1iNx3(
zfP%t=`YHiX_slxWM%2T#cpBQuK5si?skuWBqR69<ppWVx_L--)z@B3n&@*94-FI%$
zTipQV%5@}-fo8d+VmrVn-2Fj#lKivUjWvGm;P<m1*g+An(GfSkBDdO`UWeJe7oA;R
zqiwf0FI3?;1Ta6vsIPagC2<EnAq7zg5w~TRRMI8_%-F2bGb%=p@{tad^#ttLbzJ+6
z>HgV(WLhE$LyFpPJAX+-6kf8*nnqIBl{=HL0ULnXAK8bVV98~}z@k)+6oev6k<^D`
zXGIO!ZP+b<0ok0EyhxS$zH!bxhyp2S9ja}DImKHoeyw(;5)Oa~*8=U{hF6quI38@;
z4*9bEL0%H~c-bh(rWDMgEfSb;#_*#wkUg1{S+L*$T7g|267lZcJIaEcNkar`dcsyB
z?wQGrL{=4>Gv>hi#cYsb*ISi{oC{?0kk8X1p(YNQ6SnN-B<>mY<VV}n9Xk}}yuE1N
z$CUzO*}zwQ{}g;jKXJ86b5h8$&9C6e(8DKI4EY5gCXJicm`Tu6caig^2P|6Vs=H37
zW=Re)D3F6F&Y(=;)_HchIqmkZ`-{e=oJ$S;mvW0iU;p_m;G_1?ODBFeU=^upqqT1R
zcxmh2!U}rSj3A$4S9pCIDNeLjd46A*mxN?DrkA&hdskViuB@h(e@df(Cfo5)0<#DQ
z&~k>-PLdFL(^EJb4tIaOcf*U0h*T2191;7cN24lRu9+g41U-D&QW_MRK^h)+F-u~*
zoo7NN!vkZQst@otsK!DL{`3rYdoI`(#><HwKDkq~ocvW)D{otGf!7F!GvWwGkQS(H
z2xJEEX9D^k>2%TJ`dWZE>b)fZ5}e~rWn55k{7(UrQ7aP1;cTg<8kTT6d5kNRHRqR7
zDbv1=<0&o6xv;a#Rm-UoG-4C+ghZ;a9A@5l!*sdTk=H)Mh?~pXnI{n<Z9&HRR9>-u
zoTcogY4^-v2Rf47Z0eswQGQRJy&A;fiw2yFheQ#<j=yo^cE$EK3#MUXONUA2@|rAa
z#PIh)k|d4%9f0DwL(RqI;dAO((<j}5bm0>sx$igr)RgtFOEHXb1vzX<B5JNCNu^tB
zJzu5I#b#B4l2uFshd`rq`}orWjIK{1`HT(r&z)yCqdLxryOelB-dniEsU)MxT;=2%
zwp{5BIms8a_2MjBz->C{_mibz`=+JS@<r0Wt$8QRPMP(~wX)jXJ*b%g7R14rKP$F&
za>i=J%csTKx#`S<JgGtwt4ex~&d$|4I>9LWH(;F2D3l0`lbtR#E{#LYtWDoD=_drG
zrc=@+YJAxpn9)<#tvd-{*}UpMoU{1(tURt^snQYga@Y5v8i}Q9qRcdUX6>mKS<<F0
zEw=i>EsNtjd>)h@v<w}+7sT=NBHdGT%Z60=+T-T3jNC*MYl^wA-}|cg6_^Sa^U`JL
zmCG7<_)9Q5y~H{-EGX>b)eUD$JF$2BQKJ=|sYr6ail4b#%ztFg4t@G#5v9au%WWA?
ztxj@((QJhQ26}ejIgji)4WW^HKE0SJh1&KG&zqe-e2mGqLPd?}OLU%J&L%4*H*=_Q
z`CW92*3AcwfvFr9_5vIiAK_tfdbzG<RHhfzT|Xq<qhJW9WFmr~M9eJnrqbJ~z-}!Y
zi)IFYk%L><)~lPUsg5F%9ef*)lbP;A_Zw$M$SoRLf0yTvYmyyZ5^_`9i`iZUDDiZD
z|I;`t_ug=zojzPAr8{|0LvbNRYBwifOV!3MwX-&K`>2_!)M3cI+sC`y`VMM;V+}14
zB-CfpFq~gVJF(ScKy;_{AR(FD5813%BEk7cX3L%_@w3OhE_4c{${pTsn=ayH8y|NA
zs!m<tI>~(4R-MPWEk$_p5?}e=2yFP?cVFQ39>eA|W2|gB%?W&aCCX?-`}#GF*Lr$M
zjwdh_Z;k0LFx?b2iV44Z_3dgsQpGrM+yEc71AH|iJ_PK@_CH-2v(7gxYm;9&qKj5q
zr5a96!2IQ}V~AdH^`kHLS}6BuAg4n{;2M+R&G*t@MC%r@A-}3jl`-l4?<VV(?k)32
z=pv2)f08rVNC{Q_C8OF|m(tj$>r^}R@L*SLHHiZ+)v;)6h9r@@c~%~zB|-svv#p4X
zR)oiSce3F?-EG61%*)HCBY{-tNHz{y^x$foPpNsy;21e_0m1u?DInFwjD3rq%Olg5
zp}ecJB(RuofvR^bT+10bpY3ROco9|Tgks+bM%xyAWF!VDC+i*tMYZrvo@jN=LioE+
zC+G9qPBhia(^a!?uy)gH|I`r6rPZEZ&0bdi6daS!%6a4NYI`o&Z@%2QLDocTwR!*9
z*9_oTz%dZMWZ^$H-!~!h2^b|oUu{cYwe+oN`E+<7JN&j$tVjIpes)}`!whKpJhkC^
z7H$<xrVvTyb~iiuh9xf4pYyKw_Zi;w4TdKsZmHvoeRq5RRgaHi^`^f&=@~DT)8G-X
zP6^MvO6aDmq>n!f*4hqz6pM2<lb1I&<y%<X&4T&~98mA^kMuY=<lC=1*#x*)b=(&{
zn>uIyT4GniR~SFTsgKx4ps&OVVr9cB8{2ujNqy<pO~eZ5ihD=>{axc3cd4LLg_H{J
z=jr**ykzYa3!-}|<?hHrReyuFw!dsRdsF}>RszGaU**KYyvI4U;)`};LG>T&7S_~+
z9ab%C_#xXx^MY2jyUeCk($(3ii>E@c%A|kfHW|9K+TmXMybE6qcB&eRuGlK!<^GZ6
zUlR{>-pJlmdwA;U!qEF{MW==+q_d^9pvMH=p{-k5nZfp0oBIG1<&Rv*1Pggy!U@M!
z%$)qzAm%-|vX57U>1X??6s#=vH!HbqlA;-YkY=U*8L1uRu^rxpPHaaEh<t4OL?<!F
zxk3G7$eGo-W&La6oNsbY!@cz`0;$q{00(-tbD2kg#j;;tE@I_xTlxlpGpsi!)WLR?
z%zl{2cvjgcc?=DyfbpIQ;u9Z4aMLPtLn{hKt;xf5=_@WJ55F*6Ku-3J{|>MgEtU(D
z9PZF#C&>d@p!ao`?Blhxzog7JapYV&l39!0n4Yfir_P;VI*csVL{GpG$FKl))wz^3
zRJ4)ig)F_qBOjV*OF!yc>JG9RfN94+kfk)9F+Xe!q+i7k8@hjlHU2|6U_lPA>w16!
z(Vk70)Wm1nsJfM7b))(K|MJ&YAMVAbxOE=IoW?JWmOg0ey|k846M1ZBVi%l)Az{OI
z{pu>>gcFVYv4tgpRjP!ej1vuv(NZ=`oEc+P@;4YCiLh4dQGJbB@REWnogN?gz94o3
zY@EjHp|80!{Z>{t+UgYN7j3Y;rMc~}ksjkVL{<PKTQBXAo#rvYyktU7Kuz<>&@r?A
z<AZnZlAiodj0pcBx<UD@Ld+Z%L3C7HBRoc9HN<-Ch`VS3U*PHeeK@N<DmeC2HsZ%*
z3LebVD|OGSzk<W5MYL|K*Q6wq{=@?&KVVbdPbBs=N=$Nqp5Dp<80+H?q$e*>9m@2U
zoG<2L*(W-sfvn(oOz@FRIU+McO(}LS0^sjW@*M8KM)I?q$^v|Z=_t*nfBG$+L|Z0N
zpWSnx+f*5mO&&9j&p~poprd=o1c3G9tF@k`v#*x*MEc_EN+4UIB+T`|P2$$&8{O|W
ztAkRbiJqS;k8d~#>ln!W_tKqwzD@#<5|A(Lnm?L3{qf=;u`9pq&FnvQB5uguiJj}w
zhiUxv9^Ui*8ywHVFIO(|DnBxoPOrN4f?$_ff*{V@=1Fk?^c(2bY-M6?T0>wdwO<H#
z8;9=j>g#A`eop)|783fq5Ju~H{E`VCve+;6J1hpDkv%fx=9=c9s*YPSCvx*rVmBQD
zI~?6gghD{2s+SJ~F_F4+i`X!T0T<;*z~z^ZmzN7R;Nl8$s^w(u{SUZL1JaQpFQ%nB
zt3;R~v)m_yT<xmg4lC3}k>XZ3Jx*k5WtvU;3}P=dFUf?#K9_{Zi)LCi4i_zBu}>*8
zLg?X7gtqczm!K2D@Ubhn%saz3r?Dh9(#eJ1&)Ob#j3>WuGsl;Jb|$q%p@LY)>gCoa
zx2R*VugQ)CKfNNPF^J<0)vFfWd*?&2P2NjOJ9`h>jYW6<B4G4Y67u!!Irn^S*3;{B
zyHuH7=?*_)NF|Et`YFeBepUw!+=QpW>qlGtoSGM=0OGeZQ|u95HOE3N3igi6OuC^!
zB5?xx02e5RL%!%BlXTSEc=py04R*0$jl#5N(Mk~;(hdvYSloDbkJWVsz&s9Itup1Y
z61RN6^LCxhiV3k&hEgLjNd6;A-ojDSa2YdagJ8(eMMZG;B!!5DDzv>YYrDxay@Idv
zHR~V-?E;4l@QN*l$j;@sjNi2!&W&zD5ItZNd-E42U619bY@VvkbdVu30v246EkzLw
zG-rT0dg)KsL+w<_&;c?1!icFQ>xQ3fGu*3EUM^JLlEmMv+xk>gl_~I};Cutcm;N{m
zNgzz+2o{u=K+6x~XL<@6sD<RX`yIo-RCM(~RfcF@O-~sz+IL3BPlbELm7Lek?Bv`e
zF3BOs$&Cz{AR&HfMMVt#_N(Ze^g}?>62Toxgox&eE63Q$P3DOY{6pn}M#AL5aOCZ6
zNo)uY%sg;!vQK3PKsTvJWzQjL8wechXA!xz&7x|oYKO0t>o#v<)oof+=z5fT0b$%S
zuvk=&fgE58iU-UV_N0q?Q!Os@6N?J#by*s`lQ-9qHrXOz)@1-pJ&v$V8O`grQiTZA
zBb&AnQZ6lCUW<p@!I_mM-uPRTopN6*>m@6zT`Ez&wK6v3ENml>f>Z%MnR2cnzHIxB
z87GxXKt+d!2Ch<FS7i~;N_D)~r>^yFo^ol{3#b<Ps8NclfKSavpNV)q(K|*HZb4>K
zmS4@Pte3`05ngLOtgDw*{FQbCIILXuFIlJl?%yo?-#2O7PLJm`cC&b733PS<#<++l
zh!j*!wkWosz9`t5%PAPMIbk-m0xm#ShnQ}Dug51D2R@Qv2x3(TB<Izs_U*zZYk4ya
zLLm8-sUaNs@)B3agBcDZ13;d&`cpuOFw|B@v5kiqC2#j3__h>XF><RB%#$LVx+_@8
zs;lgd(AzgoJnq5^G<283octTdIlo9bAgJ$(Ss?U0S+zuTlF8Un1jJnN@Nd?w_`R9+
zxA*dLBiW0}RC#NZ=v(P9$hy$7R6Ea&_j&l12qQLD+&)%-B^ikl4_4~>WCBsfLakIa
z7{Wz+ZYA0@%XNWoo7;uWR$)!LEFxPsjEN^!R1ILQU3utN*5_G82aK*%3;!AG4611+
znEF}E#jXZl=-<M^Zw@noervocPWPcl-Ti8k+S|N+j1qjWLgKR4+J2u!NyL4<1M$4_
z1;csmW+fu03jv0(E!m=%Umme9h3YcKb_SZ^YQLGlB!&$tg!L9PJqO1#{Fe*hAy4qC
zI{Ro>xWn9Ed`VDU-|n|U;*PjgqDfr+y#fY-LG_&kkB5@l!^yaR7%D!Y;C11eqnDSe
z%`jniR0WSSn9R0r-D~nuQeeG6-jBh+Uo1HVx33B3&B4ypw4<>{H-GwsNQfHyTcLf|
z$<n2gR?U{U%V02KcCsIlD!+LJo}hwK%g`Car?mV~i(3>(MS$cEkaW}6s_{2IuI5>7
zAEQww#KY%g?-02YGAE%TLAfI+@QIfQTJT<QxH%I<_bRWw>q*eFh>vjGypIhW)wH=~
zSbZ|OW{_^)>zY+=eGNh;ltUVQw;JqPB6+)9!7dF>|Cqxq@M=}VAE&pjXb*k034B}_
z4Se;Mqu%XSlx687P_*kh@e@F8Vs>SKMwF%ga08}n3aAFIKK5Q7iiJ699$c4<PE&};
z^jB|Wc%7_IgugZb=su_|1>7GJb2M&s1RCtbQ+26u49ei57K6<MF(!T1bda-{$4M4E
zG_1!hmj}LcQ-g`1g?0+vSS=&VCFLTZ+Q01Wl`K^9$zg!(vIWjowqkv+zwbb|ETal~
zf3cck@p+W!*<{$nP?G6nrDR@EJNe|)vNl1O$!X!Z?#UX-%F%o03|je(nmIfHwe+ky
zQ0AG+fv)`sVKQFyD-2e9CRD-$p_{*ND_eT2V1tV!ibL&(E{~|g8FI#o+j)Yf)S3F^
zu|s|mj~qRi29gVCGDAI`EFGFJD@T9AtdqYH<9)OxSZ1o$v#RHv^m@u!c(X3TP*e<v
z!$aYqGs9du%Juhc?Oz(g`HtI<g^L-K`O_F$TrInas)keZ_xpr0O54Wu-l!ca!FYi0
z^|K+1DI>wSONm~Oi%6oT>&y5+$skqK?XS+Wo`gAhvKPb785=z>!6;q>gn{}MqeP*g
z+icGHc_6XN@mpL;sgSV6QC<9lYtDugd25Z?*fq_53becHLz$qKV395PuQ`QZWj3Cd
z(Pp%D>FY!>v~}t)nO|T)M6&rdt~To|#{qh))paC1hs=R@Y(>N2{R+H>DLDS0Lx!L_
zY1N?K{VtEWK4Dwj#StOUYFXI<(;VdI7EjBTzKJCqLW&UVzS)Q|Yh9nmH6n%R_UVO<
z*G#kJWHKP)ZKJ2`s}!vwVnZlVcK*)u#$^zcS<V*j5h-#8dL!L(ODdx?<<90heoEB$
znr?zx4K?VQFohvt^9+bBZp3o+{wmaNTm(q(uy${Ej*08IeAaKwt$V_1!_a$wPDg&@
z9m76Zwj};Ww3kU;Ju%lqYiZ=sX;IUc&7a8zht{NPCw(Lp*bUI-^n?HIxrOV1;Z-K~
zvekZLYVCJB*k!ve$K5Ch^D*uH5sW<y!%-opagK`&4+C9omG+cxYLSVczF5ZT;CJ^)
z?ag>#k(=cVOMhgqxb&R+P~bUG>5j2WSpJ3pC8_Vav~^66_|@~8VOZ-`nn-GO>Zx$6
z8wMT!JVtgYv8q7D|4>Lb&joS03F?pW*Xxe%Zo$mMn4+F@Y9TIYlm<qR!s4vQJBzzj
zrhYqWnDhyIt~nkr0t!~-zt{K|zH1TiWbp#lODqb*G_p((Ie7QE3c#r9y0~r3Jzl_0
zCE3A?j;<t+j0eLK7ez*l%R4RL8o+A%SwLGiy|A^tt=GIElM8?Kh*L6J9^U3G``sqI
zYA8BQRS@pY1n_gjN#3Dp)nUzZHezs9lWFP$`{~;;@olsrN$E-&Wpnam<;b&~9q@<p
zgT2v~7{-=H97QsshpdIwvvSh~v&YkY-FneoLU0o$IKKV0_EwnRMCP56;kxAzwef;)
zbyK{-a?z@bkVi{PD1o#y^&C(5i(D+}>$PeHk5_Y8)x2!p><s2bjVkVzZOzF7?HBpe
zsQLoDmzcvWVe^Ah3k98u>_+i-5OOnaDr$*kL&xcOqOL3~*7^JjL*)oy;{XvC*XOa5
zV7V!`%6h}0A))yqkLh){n^Eu+YV%JGDi4%<v{j`}gO&Q$_bUc18H)f6$5Rv*{;1bV
z;LjI|qg1{8y22orD$R6u5tow7J4Oyo(ZHQ!>(iRGV`sS)cxi@~+nnWDehMk{$<Ks6
zO9rAS4!^=ktj{NQk7Y0rL4TMmDPj?iOx+#~0+V>+uByZIyowq(3G?8#*3*++W7A3^
z8&);^Ck>Xnk3#~7WDZT5wE1wSUc){60B1a8{qCoe|0yxy=~MNge-%a@{W&X@srZ|D
zDstU6P}M$`S1Bd)+$MrY5l2^IWxAC6pPm)<>be`L7k{<<&^gpMFlKIs_qU6=ZD7n8
z&TN;0i2H>0#og)UQ(gUfaUJ^}(kxB@FS8N_d}7F}uhs-~7ltk2QLbHFMTgse8CD4k
z^Fgd4Iw*Zq5>c>U_nKRZ!0S`n8}kIff&_pWf=8bU5yQX#cVgH%k^-q0B}on<l{ND8
zsE~qcU_cO|{Le1>;qmb!RKi;wLwi!Uzc#DVO<(CJ9gWAffFpca#$p;GNH%WZ(;e&~
zOG?A(uLN-v9-Q_gX7cJXtv1-n=sZU%A-ByA-d-U$>)um`vJ$t&lU_1X(Qpnvvbwgh
z`zr6PRe!$eY`l6?lbpP_3w_5eJ&z{uK2*d*h4}7;<^?y?!d`@jXtz>8+%C4AT!kh5
zS<zR|6yov1a(-d3jE{%I^R8Ox3t@sTI~`Y$EBWemL~Jb~gShH0LsWT{-J=ueff5!+
z^=Y8{<djwL?wF^7k7V-2X9&Vp?rpDE*QjmmE>akB_M1k7{l=Q*C;4RPTyE=_krM4D
zuVk#DXZd`49YP?8k@E9*SG<sP2JcEbEj}^QYE~5ES`=5Cq;GN>GM3k_bxjk2(<-+Z
z%`#&TxDXBni8*L`efoojPTcCKgGvSabM#JvLV>mUwN)z;JcS`If61sU`<&vJa1)SC
z0%vx#94yT;ewYdW@uOuLP7mr^vH8=U&o({bPR_bNm56eBj()Z4V)Y~*yTfkhu(yjA
zB1nNHd_alq!n=b1<kkR;(y9?<+*Qf)d&{Y_;k-w`TmIRG)3{hXE0^<g(E&o<TP$k@
z@p957;3;7*%C2ASDh`jpzZ%(_U(Cia$x_?C-%oy`WBFcmf3NCUZ@HbJ_gLOViytXS
zO+`PRmIj@DdV9Hakk!fSb!Ws2)8Z2d))8}RCP9yL%$ru~_K{lfoD>UQ?pY+Mb318R
z5e6_RrCAw9-O{D+pGxiFxalB=-?r<`zF3-|UppxvKK+vC)MkFH(0Okhbj+>C!<mXJ
zAWMl7o+dI-Jq?e7^N)|ryksfh^$UGr&wO18lY-AD17kM$T?gVe$WVI8nR@;*Af7C|
zNJeTRoXAu-pjt)PT?^2OT<@r9U;&(DRZ|&%t32sotx`Pty+G*Kehg99^sM@Ab6&hG
zO$Nn;+`k+g11?vhB`1jMqh#ofO42twS&8vy6Frqo=^Iv<>Kh}Wd%v(W3|M+e(5AHJ
z6mDvE`$!j7HwesDEcEZhtIVs^af#1+o14cdLnX9so%sg5`MXL<q3PCWIjq__b3ck%
zc*-z7X@N6`q>0adOzClnoWxc$Qsk{Hx{i}ojhLH}C^FdPYFC&s+SvV&YAXzA#DZ(_
zF=k2y;c*R3=@E&TZ1<}KGnStK4Vu~pNDd=wf6mc<z(aZ5j=?!h6wsc?tpQ3Oyv_{?
zKM%{-4YXcPmx=RPe2vQ=#6$e9r->B~+tb5TZx<5paEM8dJLmod_q`-O1{>K^UDtuO
z$PlHvWh&KyLQCNuzJ4QjX*Q&~$#A1%^`P0RlXlDe*}wuW`Vo!w7?21rzltEBlOz+D
zM_QJHwuaGpJCXo81$+81m@5(XsB@R?j(_b#iK=dYUYUZ-v4tCNHdhs8fx)~*2^Bxc
zJ|@fO7nHyCRcrXZbR~QJrNNOZ@%9;&!7vZ*VOM1S$ck<ICCwPG_h+AE5d3{BAQ>&!
zCRUOsUMvEL;9L#-0a#xYna;COL3b=Z-I(O^Td^GdWv``G9SyoQXve4OrnDN?$5><K
z`xwV6<d-|E=3bU3U%IrWua8cZ;GB*-%hi8R;hDH1zjk!}NtN)uRWjH+nV+3`So4?=
zH!lVW&4`Fu?~BXoWB#g?_{CqCU0*@xxJI?cG-fN9HzR2O3L6u*pQjqIC5D)(bJofU
zJ2fLuA>s5{a`0;Ny^DfQ+bS3;)TzW0DoP^~eCAEGQ*s1X8qcGzE%RSd^`+L%tK#I%
zTqJIqfw+nzVe1n$ZiBZdvf;(|J`Xe6%J%cbnr9oRp=uVl;l&U7cl{l?z8%i*Ki6l(
z5vjqkmEX$jmd;WS(A+JuBO+1YyNDZugxL`+%^!{lal(r4PSFq;io%FEg<Ko84!Bk9
z>ImA_PZ(2u+BPiHKjYcogNx38kW?(ZBY1s55_s5g(R^aaYacKo+|r4_X+J0yW_LVc
z)joco+jOL&Lu~nbe$pJ;X;N2tdmG|RXy>DA7X(AS?$3&2cLUB+t*nyVB59Gc+L|{P
zBqK=Rqpd5Jq&dIA>G=i6x@N?w-$mdz@e|H}-H5ZPhDxZYK&m5>qLOkCGq}&VZAGuu
z=$npiW)Q*hd8wzkDX(o~Kuh^0H3JT>|CM;m(l;H*FrQuTW6hx~^wf%^qlTsxv1DU*
zix99mEekt8%00~oe{rCAp<p44(6iu?vtnVaA4%bX#*4GRF}1E!i5@Oa6|t75xB0p~
zWs6y)*1bdie=x`WvrD((Jq6M%HkKjw5$VMxAp*$S&8oP(|2tPa8OA?N)GjfV^{*qh
z;R8xm!jwUMJHvuqPHU05=MQ8;T}VqjoD0^AkRGeIAn+z$dZ;tyW6q+<%||UXaC9u1
zVyB+hqUCWuMll`|$hJ~0ssQ4Nb32tq+j#uG$>g!N-<#oa*%IA)t_QxFxHz-UHZCfy
z(($*KYb@jeTBI>OcAP!Iv3j`sc)&sgN@0~->5j+a!BphqZE@YOQkr6#p|AY~<wOgM
z9nx~uyl&gYrI$GRnV%kZL4ORnFG#aUMC|*!OFD#~-D7puK67SN>Y>2D9OH!Hr>Hl3
zZ4g?0^(FJTE<c~6E20C((y&-&2q$XR+v{pX4+a89Arf1bshp|&L|16_E>av!%I~({
z|Mwu7O@*>kc1nDx^`6;kq~odhUT#0)#A;R)vxs<mKmoiP*0ce!VI~s0cdr&pKOb<m
zPlaeddCg7s{gnhkYOC=~?x$DQR5=<dj55#DoI;%?^+@Z=xw}Xd40RhK4~uLgciZ``
zJF6=Fe!FOWekAFNIsX>V;D9N)<0YTR?IIU~=$mNZt)9mhv8|V@@CU_h9AZ2K2M>>w
zu|PIh-s4}g|K$S0K71fh1Cp(OfF25`7(PefCmf4R38UMrdp~<Z|CnYZ7Ubl@%!4rs
z8F#aUyh3E?kx*}%R?bQ%@j?*(CM9X={+>jJe(_7TV8J~hLWA7=jRp$q!3Df}OZ}hS
zBidVestu+0*_oFe7Wc7G(u-6aR@<K)H$keqS?el~(^=Pnz)em&5$~Iqs_{L&s0<!I
zbyo=_+s_QuPgeJenXf!c){Q!7Qi(k>+X^JyzpBM9pp$P);_z?&0-$B-{jd94C6(5n
z-~3bh+VBZTx!v^m*CR%mY?^VwFU1M4(k}T5-?g@X<&VG!H`UZTkCc-ZhSQYZMF&pz
z{3q}DNQXiM5f70<Phh7T<N4FM_IS~u)yQ?>>~rwQhYFFZb#yEQ*3b`CTWf0ewbN^0
z3vsxBA&~GF7u)YzP`-(i2Y6L08DIG$0e(3BNpU<5^nmyaSb>ft6ui!;JRP((LYoYF
zmWrTI1`N_=3|pzArJejYJ(l~?-sik2vGMcM;>-%`6A;;Mei<^Mm0v#SuiO9=_Zf{n
zCjJ?QP<C^Pvr(&mxD~iru9;F$(4#<Vtr<ciH|48db5dTd-!Fc8CE~Il|8es{Kgb%9
zIYlIMl2x^>j@bT|39eqtA;cp3Vkw%7AAWQHARiMO!aJ~sAz!OyXK@>qr%Hu=ABDv4
zz1yo~;~uyjWDfW}C(amz^KFScpq}tUil~3_<0_oN%+H?FMzf#|GE%G$GYe;l1l{uv
z^TeHoo!?4%kV1XQ9V9z+5D93pZ|Gv=L-l*bU+Cf#1>6}i7<=r(KZ%24A_*y?bN#!x
z5AOy>3K9Ff02VuSFYbRAxpXn^&T;VY$RzktjVlp(QYd7a0_SqR3ikP$L=cas@iKQ$
z+1#hVDkHd5|B<@i(PVmS^XyoVjU!tnr^?JLS_6~%93~DR#%0G|l2YEr*Mo6kghgoC
z;dX)j$)!P>i7O{18VTrHsT97r;jn+;Q$B~V8~R-IEFabEwGiyQhs*wIQP)CfE#dhH
z@#BE@ZtSpka{Es`ft8-wb95gH>1B?uwm<I~WPT!s>Gr|M4uSd9-%mJ(gz#DSk2fOk
zfn(GXfzYxm+!t8Z{Zvt78CjUD-T9U=o$Q&ROnYye?UaT^vybMaiLo97af#0l_S2kX
zg6`;}&lrb4h%bJJtU+;XT|W4&)QAF@ohQFB;oks;2pOa?LwH?kd}mc400HSKSAWZV
z5SI>H$VPNFqO3x0%VRCYkcW%A?W+bF6ZO%4DMh%CvxzeO(=UCT^)b2gz&rRY$dnQD
z>N7r*Q?UCOO_8DK^V!n6wOt_@i6eQj3(-PUIe$sQ4C|NEgDJ8hd3JQi>dcT<jpDbL
zwuXM1ZxB<H#o;!O!PMAWUR4VU(4u)9*052-oHOF&v9ne-RER#kld`u_8_ST_v{+Z=
zvSe8DEAdcAqJOOL_slCpTTVFj_`WIPcUBiS&Z+wwlY}?P<)Qnh-qqdV6d49j_w8RA
zo5r&ykHw@K^DU{90v?RQ6q%-;o<%}L#?$Dp;ZSX}WyzV~+!3bl))d|dItcD~PO%}i
zv)qo4gBVKsEmo}kY<!AY`kD1O!x5s==D4u!J`s(9dMIrMET`&3`J&VHD3PBd2Fi&z
z`-|gZ6cDkIPq^K(IJm)=4jN_it)Zyq^;o9xub-=bqa)vSslZ-6Nk=`6nPFE>LjB--
zIV@6GePG^hL~sgrNU^grKpye9Lh8;T6>^1oJo8GG+y3(oo|B5dfi$~~QZ&@*Ts5BP
zpU3%R?m3qulA;FGF=iDU&3&}!Y13r!&}fQY13W(HBfp%kdMgCva`36BmF~JQ9NDL$
zj6V93!Oh)7j}me6BMkvSi$W_DCEK^z!iJ)&$iq~>;L@bs!}~1%4C`OCubf@z6`!Y!
z75qW*R4R=nZ#DIwj8F2s(qhF+>GKn<T$svA`bV+>WvjO)h}wY90O4FzD{`djhe&kr
z?R9yTP5RDFd_;$k8D?J0j}MXOc<2T%v5GocdeBXIkl#a}g)!LEa!nJ7Qm#cFCBd)o
zI*s5)?ogf-*;tKtuYbnsXFl}%n>klnQ+&i${i)-{-=Lxp%Vt)9IRiwmhZ{EbWILW+
z;kNnkH5}~k=Z28z>xe`ux<VsjQ*OY~c+>}FI?tJh-!#rP{y3T(zN7Xe&B#DPRl6No
zs(2usj7388eY;+vh{mK^Yvdtj?y2<o@nb}WEDaZNw%zodP3_sNN~8EXUY4p4w&yuS
zi#!c5*`C03et~|TB8SBGd(#)n<Ojp6x44$$SWcmQ$$L1SE`{msZsKj8Z*3l~Nd!?P
z@U3I-?PlF~m@7EVz@I*T^^m+PwYfJMQq3-I^4@=7xQ9qyD8eErLr_jI9cA~7HOCJd
z$G)<<zkRCDVzRG^)tYba&)+nk&IYI+Lny*_<p*c(qR^i6*g2l>J}imA=VCS^T#b+l
zdF{+R({?<CZ}n(2x8qo_{G8mA=&{m53UVcG&cg9edol>n)FQSkCmL-$c#+_T``)2j
z`F>IwyDb2V+3?-^2k4*&hNyJrn3ObQ2=|Xvv1U0X-Ps<RlNe%}@;5HVUG9GgH1{}U
z3YJ>f9(xt!Dkhocj+Fq)t89RAq)$YM7*g=X=r{vpwoyJ-!%#46u`*mcCWr`LzeAK+
zGF-dox+3p!vcIdNf@{gV0)Ox`9~B7|RH@MYtZ{po`L0*UKqrJ<CfK~^yfQGJ&Cf{5
z4f7p!mL1Yz+P+^q&psgigw!jlvK~2t@x5Qk4q_4@oyMk|$1*w*lRd=*xm{T%0C!o5
zqF=|iXbKs+APmPgviLE9V`%$$uAoo~B<E;oC+6}_C{#d3ppu&O|A^oHe^{3P2U-WR
z?<#WIpwPp2okIl!lg4nc;Q;nAvD|a`9TVp6yscw%l*%k^=+m-+%8JA%_h^#tHXiAy
zxPVqVTtnvgClqSgXzeuxTpZA!Q>ZrV`q?5D8VF*4_`;A^heV2Ar7ShK(c9j=Y_<d~
z+$?!{q34P;TUbMMlD8Ls0j-m&kJ!&NgD4&E4y|AE4EHqC`aeU2wPP0p@stcWv@x6?
zV|>7nF`qfLjD*c3V#9uFT(DeuxI`&;x!up|IN9$Ry@5AEw7Uzu=`)$r;rVpuRbsx|
zuhJKWHfvcgms(Z!>zBPXAs(KvaW4`E#&J*Z5IX6o)gBU3P@bMU0ezHC(}e&zxkKGb
z?~bGScv`c^W&V=Ev&iz@I0puY^GbzgznYTJvae#(SZcKHu$jF4{$cLB8f~JOkJMS`
z(}|(q^Wd?ocdZt$-%|*Iq3`q24308JspoNVX{E-hb;D#9pVd30Jnek4uaFkS41pY!
zT=?k-ZF<!I?RHKf)4k_p>+Jz1EyQ%l{A_i>`RSf6*B7+j5?x6~&Blh`Ut6U}gNu@u
zF;f^GqyYqP+&Fap)uzX-(gmLzzCXRpmtpjkl;#4Pi;7-6x_lOQ9ACEHU(Ft`dW_T`
zE=U)4k&tF1i;8LdU#>HkmdN>>U*i0Iqt{{RqxwM3`B3g*69<cqVE29&@!%JHA-2O<
z5#bjIRGIpOiYZ9xbkPBMLRTPV!b@@KA%K#L;i6^8T;z2+rs?f&xVxxXMC4g??n9RH
z1u+49=t`n=d*@NlQjU+~=yDI76Pu&a>`vbhVI|xq@Oe)z6(unNd)4K`?%`pbkH=yR
zVccq!LX5EUK4wM<fGOXWJJaKQRF$!#J=<j=@mmG=>wWdRNlC}lSlb8kL{W<7w}1YV
z01MdB<pS(Z+g#uiX$w~f{Ihf@d@uG<DfLCqVk^YSqaG;!`lVsF?)2qz%aq@obl3BW
z>aV1T>dEa%f-4CY9|`(>rjGqXR=i8dBVO?<ot^eGKNWq~k*upkjTUvgmNLJ&ma~{`
z|EgU&jPNBS5M*t=FywqIGet{9g`0J~(dKU|!=hF5Yhmz{bV+d6c+ynP1gw9%)m)Ac
zqOsLxb9woXp}jy0v`JnkzSXOYcudqE4@-uup-(L?rJ>=7O4<DAokOv)Pp5GK6xpcs
zrq<T}gS)0$4c0DBPK3vTaK1p#2!AK-;^aoQgNzZ|&(*GAX)0R~I~6C?(=CK3R?i-K
zXGiCYv;w!oo`d-6AQ6WWJr#W(idvpwP(c3?(>FJKrVSBFTCgt=nD63z+rK;vnQ4;7
zmX{5#bd;5QORitRN)9N7Rr+oBdvRU%h1acay5yclgREV?4bGsFhzQK9#r)YZxgJ|I
z6co>svcKC!!8_^6$qJbJi?E~3#(n5WsT>t(Xx7axiHZZ#$1h=XOL~`59W&yLVOog;
zCN%cl8Z13OxY9p!YqG+FKR>JM!XS!ms!xt|a9X(%;hGePHJ$UJz_-XJtn=WfoDzez
z6mtZ{^?nXT6QJ6!Hpec5?nTL7F#I1+nW~w6kMj8tf{m|tzoFuf3;8{ryv#VZ&J-T@
zE*;%38G2q1l%kW3mz9@OLI%daPp7#<vBVcUlxCv{;&kt;Xo8w`1EX*mkno1pDuOeE
zt1ipRAdAiEf>=*=?K_mA16drZ=BL&aOepEL1yZG4*6rKokw~^H?<X-zCx+LH-+K<e
zS3-3V@uXBVq^72ZL1=vyk*Nv1oj^n-5EfhiL`^-twFT~c&LQJ-e%K$u@b;A1RtsJe
zPzS#K6)|F}54@a4B=<!Q^|>IFgZzAj8a8R3-YO~#S{X%kfz|ag!=-xm%`@MP=SOTC
zF~LSmsgh>NjwjlQ>=I%-zwV%wL<C7g{<Ky>o1{W2VKP;fr|y0Y54hm^=5|O2m`2mu
zQu86Nrru;Z5BSz;P_vYHTiVNpEH?bd%*F?Nalh$PEh#N6AsPC9>3=yi1QvEbo4@~z
zgYAt$#2KKS=d>1nf0rbl@S5#mkRnUNVJ@G=Wy8Mbx->seLp$hqGqUCgCn5q`bLfIg
zUb5K1+Q)~!mYtT3Fy$uyxTHN_&Kcqe;Ry&50fh~qfwG@Ghcq&p5p^x}9KKYp*ZUvz
ze?a2u?P~JdV*7$UWyoChRW3th%nIk*;x9q!O5(H?I#y8~6od2yH{Ciif1^<jjH!S>
z8WxsO&JT`n&s#LI#o)*5PwebjwWwdXDe2R*{r-3n54LF4A4QfBT{T>GfFm_9hB<|O
z9(Hd)r<ZLGPibs@5ln)kAF4ts{a^Mn^CqJS{717c5|uiwu<R{<1W+`=#Q)SZ_$%$@
zMbiS`4Q{!Ub<69T9pXdk&z~bMw^r}B<GntfI%?2dzJxj`>anyofHkL0i^U|RIEVK2
zB1^S#UK8vLjvozP7|`ai!zUoHV^0_0Q=#7mU*R=E77zSujlg9t5|&cO=Cnzo7-gzV
ze{%8y9H5mZ2di(jSfvT+sOy}pWK)bOl0$P2YXsKIuDbUD^X7&naS}m{@AHQRT27+R
z%y5AC%)q#1(SM)a+t7*aN*+xhNqd{TG0`E)4cMfys^CUWl%h^ZM`iJfMz~&5sTJ{2
zqM)kJybh^<Q!KpC`d8mqAIu`V?kc#-eYqt3=pU{a|4i)Redy$Zn)IooFs)Oz3P`g;
z(BzQt>Jj_K_XMXP2)A@2D$OB4n2W-L=wsXeoSpxjF^tUH<#rtDSu}@nWSl6lg6h#Y
z;)l&S_CyPEkSvQu=}Xb>pTej}N`6q-iC+)ElvRrs)TcFUNU*6u&ZCA2z@%`Z3R#FS
z0hI{DA3eRZjbg^R@p&9IM&F%uJ6QMbLUALKLI|ZIADtilm@Ln2#Z9yw+s_fsieTET
zRVk0W@_)9IY)w;fCGV9Yg#GfNMq!@^4)zVc)1r0d2+!<%YbZxh@bHq@$d*)9?M-4Z
z^%S&se<-tCt5Rh6{U1;<oTNZImoLtzFc;f-)r*52S@ipNuYHqYyXW7>YmO$9MRPa+
zl1~dR$|Iz@{mfhz?<GSUtH@OOtLj&|Lj&l4C$A>OT0GQ`F3n@jP5>^f1mAuy*R0uJ
zjx$u2{*ap*6?Pf=`tZlhFf1HVS3hw`DU0{yr}Tl)!*VoamN4j`%JDt{w4sv4HCfK!
ztgO%xeD;RTm`LYHZZ`jh5DIeSwQexqY%g_fy|}rvf&i&y6NiC@%I!h8$hr)ZgM_X6
z$K@jV^ggG~w^NZZ;TP+HAFq6DKYP>$KhS26b{A<Cr$P+U%O1HuLgfp%tBa7$1fQAQ
z|HaMHaypt-cx8TjgwBvq*mtq$YZq)K38$r~wKcZWgk6{Q0<C54Pd1n9-Q=8{sB{iG
z9Ju2u2xpM<*Y7-m3gb2@nh>IPA^AtGW<3wnxN|ZYxJ6I_-G$$~$e4n;oSy5)WhXLK
zCoeO!)ZLseb?0qMmiovKs3csR+TDlvDH<%=4W@TB`$<5y?a8jUUh!ODT}Cs1%jR}g
z^~+J@?F<&6gGwy#1yZ7H^L7-9p~#h|RBfcEkccpLh-fu{k{ck(NpX(*zc<&o;oyv_
zrzyHt)THRYV>dK9{e2QzXh5qL0#nS3d~rDoec4UUd`J`siHQLW{m9m-HG!>bbp{PK
zHyZ$f!k;s&I9r>x-aO|@27(c2$OTf?)G-{6XNzsfye)rJS0_-&cz_`EC;fwjT#&ce
zt9Ziv`EKrVxRR22nfyB(N**5t?LWD{CcP+_WQ49MamEZDJKSL$%k&;2<U7qMN#7C$
z*Y8)YcIl90s)RJ0ct6MUe&|cVja0|ZGf3C%DFp)?oHQ0ARq86+4>LVF(e~qPIhT}d
zR|FF3uT}FJTRewmB~ye$GnXHC<Q+cTs-EEUbGkX|aU{HAVNMeOEabgHY+&KY&2anI
z>9c3U#_rRq-+KZWBy=7~#>PF{W|!+;d0rraOW&Jtk(c|Y!9=>aK+VTVEKPZ@;rKw8
zL&ac<g9Mt~%!w2lISp*M<G}#-Rbb}df2SAu??@k$`j031U$y^39btQfl}U#p277+i
z?&#Zb6qUcdm&D+<gK4{jWSMcpim2fbpR)Kpv@1BCJm(ZrPtW!lkOj#W=xBSVhFo47
znM;6MJL@<WxA|W<Dh-26uD5GllsEl(J=U)Rl`)awYd+yNUXF!@HoFq4n=c%^Vm|4$
zIFYAybZ(;1Dl4n4K|Rp6h)2X-qE#PG9dcCIEc$-JM9MSb&a{{U29>>R2Y1bIe*PIT
zV4@-^30cKcn*OxC9RR^;q)r?d=&nD1h}cC!>RHu;DD;u(_rlNw%qO$@e{*yG>U#5L
zI9B&fLfMs%)l=}{K8f9i6GyDo`rVJl&yp(Z7*(k;Eo;7D^y0>{jc~s}60PU}SrS%x
zA?)iYvirg#pXotrPK`aSH>MJh&VUtd@Y)C4Wk8?W_U{chl$9b#zYX>^>&GDz>ZODA
z!R4WT#caGlETj<G=3-?0T{G>Kxzco5qfGyi_X~1SO__p$0Z?DR?`kKwCnSsG5G+sg
zxNWm&`}e5u9iUxyr_*{D?f?FKemqA`i8LGgKX6H*`!G=U%Zd6oHH>`0c)1<Bq%nvg
z{aNtloGwwp5nzh0Ip^*}+t&w`{nEjj(lGOJ{9CB})5M?86HGm}A!qTW>!x?p(2)My
zfj&!_rVT~Y+vHI6ny@~&G%pX1c?gj#t;I%@*2$Y`3tO{5uk?9A^z9(i>~K<ly^ij*
z4`a*oARXS&_u4Nycz%fo7caSgXb5tXZT?!wfYdx*GX-=}QL;U=)Y)jmRP@*bJIkc#
zNOL1X-5?ml39<FUy|RMdU6e0Q1T1M>hdEeN!t532gP4Dk#){%AD9@VX4T*@(az7Fy
zdqi5{gbBYt+WyjOHae)k7H^0RTfd#$$*rOgrD^|^C{+yLKuQoxK-|x49;5>hMR~yX
zH4?+TkDV;&?(P<B`CM{Em=5+!1ks8WkeKj))*%!2CFwN2Jy0;a7{|<xh(b}4mHo(0
z4Vv-?HX3$4GHEg*2f6xex1nu)k1*Yg`q|4ukcdM_c682gdwY#?rA9=_b52xnh)&wX
zV(0#>DPlJ)b4svXvF>v9{U6@Tbc&cG6L#bF*RnQ@_zb+^s*vc1w3V^)1M5mZAo4-g
z*$jmFDW|9Vc@KsdLnKb$GwSgtBzrLsWS!64AR1{FHb!##ibeN)=qSY4syz=uqSe3`
zGvXHUN63I5$b0r1!kg@~5ptqSm1_^@O=D@!w<!+o*FC!fxt(uM1rT6D{7p*-e!Us<
zh_tx_fyM@=guwiOczX<2tw$K1*X;tAz)ofFVxpTV;mw@sz;3;Yauuwmo&Hjh!X*RB
zb7^Ya$y_0OO>;#j?QeOx;D0%lt(EoyZugz*W=shPU9I~@I5g_R%Fs|D<uu0QMeVv?
z9C`5k{14KyYYL1(EApdAfU+Wg@k0-Sf}j+CiApx=-PRA?RN6Z2ia8N>2W-#I?FanS
zZ@FFSvkxri3$nYT@Ei4E%~I<6hAWlz2{d_+n0b&jkFFx?-kTs+Vz8`7v~N2emzy;b
z8YYCd+6X+zF+P~<SEP_aa6}(LxV!SSX2v{<l@=0IL1_KpN-`@X+H%HkDOV_ZA_oC8
zG{aqS-PSRjM#rk0|99e&^YFs{G-|fA<G8#Ne24;4)fNS(S0Y19%64$DKC*sEraZeM
z1j#-=Zd<KFqz>l3BCP5EPlRSX#)E+beE@{w1;c<&Ef(KsjDWpjYfd6Fj{{waiv+6A
zH`r1R>6^zYvZL?RqPWMCNy4vd%%rD<oCmZ!jb-syeObu5PC_G9Q41`7621JGlD1H<
zz}m4YSYOBJ&5bevtRiV+u6e*ekXaAGv<S@tj2{%|Fs<@n<hw$*f6wu7Y{q`%wCpbW
zTF7F^0+}#BDYOZU&;?OsN9#Hh;^6I8+zGm#F^@eQmqj+*pS<)s2k8y9U3T)D^kX*Y
zZ~$K-Am{*jq;Mct;37nK($iLbxf@t}wkB@697?f-MM?^-)8cx(I+954dtB@h6!bzT
z|M4p(+_7SPUn@!GPj+gWoPu8`t6}7t_T=neD2ok0{O<fuT<F^w@AAl+RkLqe!yxaf
zVd6k^_qrbO0wr_(3mE#TI%nlBimxKh@osiByD*CggAq%+PnQhe;i|*YL22?OkvP)!
zTb*$fNV{aX13S(w{d|dy8|E%H%>ueDaK&RXr%moQ=Gn5hZQ7>)gkA)_`MZ~xc)KlD
zT7~#KcHWLnz4R^;76g1k5zUJ3g@F9kfH`_BR8c0LOMX=i6;lshvJQvY<nq=HZ91eI
z-GYK%W_%{*ack4Up0%}T$WghCd)~yLbDK*lAs)bHn?tt+H(GL1^Jh)D#{DM2Sek6*
z<k6l2hh~lIx#HnY^SXDk`Ep)o4|FF*s&d_><3(iMUwNOp&@laOUYcQt>l@YX>}o4K
zS2gNh^5jo4h*(#DcPQTm%#t-$5q_}Fr;TgyG$hodj@?9M>{+nHlCwGP{sQ2Aa}%jv
ztF(EU_<-`Sga|wyVst96Klz0530La{d$g`cNmW(g*z0{ACJH<Vga^5l=sSh#-AsNH
z_mBjFP?1Ssz}9}ZTa0^=H@B_G<0PG2O7+&>d2yPM8JwDz0^Z1f%XfS#OsnkO;YD&L
zrJ(`&3QsL*e%13J3Ki4P(D>)VW5Nhe+1B!tC^nRpmsOX!c-P0sU7*{7uw#Z&b2^od
z8QRp9OXg0`Sysa0i}h&lRfi)^odh6h*2Wn~!e8vIbRqa-%t6gI)KAQf@s!^N9fTKb
z5Savq_z}n-bcqtps&$!vjx*_)60g+qo3C($%00Ak+#J;jqsqaue8KnD%J&<QzKPy^
zn(@e|use@s(hBZNo{tLFu`FC=Q`3ORmO~bCOA0@u-H4vACbgo}nEU!H@RH#R(eP#p
zJ+jdY8Cf5NsA-^QC5jFyF-INlc&Rc<C_6Zedm{)rXoF%Q_j?AZzysuwK*I70tDX6E
zkjAX>We*pH;i!0A{ABMOc`!@%e9;#)?a%kGa#ic1d(Vk7-SnwaMw~`LU=G(vJt_ny
zex?zlx0pV&qfPL#j_E(gQqBOLFH}_ZH8mj+m5eqOVSL6_jyzo;>Q9r4$N(L#=<vH-
zg{{*oJn(I;)9^lknH{3r5*Pvmywob;yh}9xKyQl9%fcPwc9%0Tp6|m!37f$Fd+P6u
z-?lz|HyX0c`gb4;ENS?^%FWeT`uHb^V%F+QeVk6YxKShU-85S=uG5Ayameq3j5IsU
zxJa|yMw&qyH#D-*1C`LC*+~Cw(1e=gfE18@;~ePv2#181`TsWyREN6|BRD_?EG@h%
zetusM-l%1(ZW_VIwln^2#HOQ->}Yg=H?D>_>3_@+r+-dezwAK&mkS8+q<Pv#nq|ow
zbIlbM(fRj*oqum5A||F@rEvE4Hd>>4b9|iq@~||E-_2Jo7$QqKIO4>NdxBc@SW?U8
zt%=jM-Qi*0Vh9sWObgvFC;46J!&mr?L`|HXw@N;pZTUamfFNo>=DB6Una@qnTjL@9
zpF%fh!*2h6i9jKPv$eC+_~j9+>;3fw8Rr6VAWPV9_fy<>onhvOi}^Aaid0I<!0GNC
z7{+_=Ps7E^^y{6k7E!xL_Xxa`s|~ta{_k?jPwOsm1A`CN{4ff0kd)^vn|+-gx_nee
zykU?tvc`UYWTE>x|5a0$U!J<R-SSl*j?(!1ZuI5z!#V8WUVkJ8r-R`4VNp*~kvE|7
zV9BV+>oJ3HE|1quFj4FS<h;5dl5A+$Oj)_R*0eOo|IGnnHNckPdSM8ECvw=uC==7h
zQ{2}6u;3>FK->F|KaXESVSw4PZLSlF4UKG{Dh>5qk09%P`xYT}J~)Q><t>#-HE|cA
zw4XS(WLwnGvIXd);bnGiGbMy%N0!F-AO^f1Re1i2VScf0MS^}9%Om^&=_=U6>R?=Y
z!TR&Ov_wc!iUIWP7f#Nj?H*I9rS|@>U$1EszfxoW_{jFST)V!y>iz2ll!}IkP6AyS
z@<)bx*D))+m|m7P@!%kboZA+2xjZk^Ua=vD!h>Ynne8XjRjn;|Q&?MFvfNMQ8l#Jl
zDC>nmQpCB0=XD;xySCiVXy{kjoc!n0wH)v?$!de>z+DeDQNsLU31*XrT~yB+6!`r3
z%e_iox2lKk84k4}wDd3R_iJXlo$qF25cPLH$`5=qVLM(4@V9H2HBt|<`kKWy8YE5b
zhlR4anWKFB3MLmPyt^n)-9+$<`SAk^5HgroL*c*W*YvXHkP7I>!T+K{z@i7$#;pAZ
z#T|MN*MQvT4ERG%c)@z&6)aNNGmSW#h=lrp=!1^p9gO^&a~HvuV-G7I+Nk`c#>58j
zbpR5efP}drt@8wI?}hUDHSnPxOXr{O3+NbS5-f@9M@aP(jqWxF#l3%in_2|J_~w-q
zj9TF@%yS1Q$a-^cZW?UMp1Q^=iM*Z+#tL{t6lO1KSE?+Hn;Y6%<2flrLyWP$lm4&f
z-a4x5t>6BY4y8kBkPhjR1`(7l>6Y&9PEiS^yE~;pQo6glySwkgy`TN;ed3-o?)}{{
z?!OLX@Qu$}>pkad&P8^xSEUuuP7`^(+y;3#2()V*^TO>R(Swh|*fLalsCUpv1RNHe
z!YpexD9SL!Zj&WA4_5=BK&IeWaQjBwITzwJ5E}Gs=_Emf{3GpOyPg0)n3mX;+Ml?5
zJkPBYsqb3o$gk9FVU@{EWET%>90(54W>zVxG#ixrTKq}f*E;RtG`q^ycA)-de*@N+
ziH!|GD4fNB=eVCcS-^vWmiFfIXe1@TYQ$wdR*A@Y6(&(iOI`hZ&ar24CM=5G>vjKU
zEIe#F<<oXNA}@J<^J%Bfr;>@+hqd-IO-NT;*<)5wF)>s?f#c>AgRzb?Qqlk#yDZu2
zQ+4(%?vlu!X_pEaVk1?g!?X3H>NAtX=vm*~qcARl!#FOsPt+@3P`X2Xfh*}YkZWs^
zhlu0kWr&FILtXRr=~1t4Y?{1;vxZyVAVw0q@!uXYDMv_<JMFYaGI{P4^1B14NMdcO
z?D-NK1;z5bYc`$6wev`oFcqePf&vC~%?#S1v%#0hpi!l<(TPzu0l`Uq8x;jbgj#k}
zxqx?6_>p{zTY=fW6krGhs>b=3(C=cxw);5xKL%+kD*vTr6Q^uJVa)K+Ic<X2ex2zX
zcvm|9`)YK)=>fhQ?#_i@{-K}R(lmaS5@T;26yMG)XnJ+7*Ex?YuY4>mlSVUaojw|T
zV=E_jota6_f7m2pRV4`h^r^Tu)+rMm-DUlSrmsOf$H8msWivg!$%Th|M7?<}*FEP?
zPDPsavEUnk41L|4pjW4y-t}wJ+@Q?A8V*P~!tY6(d2Mw?<b+sXZ~`s{>gL=`Q`V{E
z;VS};I=wz?EQvkTD-uij@F-T#aZW0wM;&_tFQ$>gFmyvxVvlHJFY^*){xjg9-p6>}
z?eH@+T&`wGei7s#K|pR`8TCAD_Dps^|EA?}h1ebT((__&Am#9UhRB!LrxJZ5MZ6{X
z;Z~Z8PO}bi)dtMEO=^KC@Zb3tC*It}v8iQC;FMX+UC7M;ycj&+y8&kF9uVNe#Dboo
zn91xo@g!x>LP2-yW(K)hEsrD*nHHBUK+uW$nsSQUHVw1hDb}RI(uu}ZqG1bU5z$fD
zRFOu)<$o!gw!lC`Q<ar%yUoZ(fEV6AbAeZ6FZSc4Q2sJD))=JO!a+(y^Q?Cm8!BM2
zWbxq5vk2nO!~nzA5(i^UsCTHB4Q_<We#nXQ${*5gY%vK=7XU0XtTWjONXA#2G3pLn
zgPuodc>lJYVZffQlir7^CGj{v+2L$*X1IFUOxX`ciFl7jgLD9HLu+8==KG5$$&OR@
zlV;f7lWLn*%1HK{Pho(tJQ<l1=@vAsyG1Yk`f(J)N9J}rR8Li>dD`R*G707}&h4o^
zi##j5nEx+^shOuNGF=8NWkflC;4kx1MD(+nmMONR`lYiCZ{wPJsp~zeL@cY?s~Yt=
z=<x?M5LAujq96Tp4((C_L(D(sDV11;e;g3Pq*_gKd_%fQnf`5mTfo@8KM#T~|D!hN
z4^7>1jrlcG`^mJz%A@$sWYI+)*ivJt?()!T1cZN^sujh&jwP$QlP+D3>~tW^*N_6!
zNE0#CkSNifbHnYr2|BRFgYIIvvVKaYaWm-kPvq5vE?PG*_V{N;y6gU`#@tw)@{#<B
z+&n(1nF<iePj7z#d2yrSXR*Q<Kug28F)I5x-^%jPb7X4uenxU-E*p6E{{1HKch7Xc
zc?~$46baLP3PsSWclJ~~pBlaY!8ebXlfZ4in(0%@;Gvu?$I%}X16qRw_m!ug|5Ps{
zKhYJOYIu_kXu|8Wtufu5<^+}!dyulFC}U%y|6VPuNW6&mwcRq>{C+vG0&MxKA(`$r
zDzffl@kP0&GA=!8W*>5k`vst{sqA7C@Ei&Eu?TO6M=lx<Z))m)`GR)|l+oO*%5B+y
zc*(oQ99(&nwI`iWa~digeS44{jQwS;Ve8|+e#++BPsb(}8l4kFgrgCAW-RC;KLP3O
z_8ws<Y>Aq_*alT1g%sWfZlu9mI1iRUuPGIzyge>2YuHt%34|>f6Z>fWdO7JHH!*RF
z$7-aHe~KJ6#@tpd$*yMS{P*>KoSB1yBkaGBkpoH5o(gik4;7_BQG9L)-rZUZAYDuv
z5rYQ>MZQXT%b~)N6F(;sJk2mAk>Z^$p2uf-=O*Ck>C}J%6DH{o!SvtlPgrhVYJWXA
zTkHbSej#0zOoFwD|2kOLBDn>jm=BeieGQkkBVdnzOMH&ag%8f8{X*WvRnH%7mMiqk
zwqNYMpTKt(6gXXKR}0lBV0)c>CGFhvk2S=vCF^IIm-I#n$U-W4{Pb8E9sWM%|1+)w
z-p_)jUZ?0^OET(3VU%IZ0)U07XM_qeV`o7o+Yb3Z@YI-s5n>MZ(e|O}X*U995Nfe7
z5x!t8)*$&5CHmmsk&(%6^0%^G?ye_J>g>TEU;4U7CIP2<f`#*SQis0%2k^)4H?ccV
zV;*Kpzfe_{OC4vQFF#sk@;x~guRIv2@fbNp*y?is7+cz9#v5Vl@USHB2UEEJVlz8|
zm8-Bktu`$$>9H3}td?tRg&4oE|8RkPr|qSV$wM{3=oJ*k)}BW?I(2aH-!?B<Jt|3R
z5&+d@?X*0P0H`iBT_4CYC`_;m=pW?BF!J+eEFX8{ey-gn=r~7Ad`%4nT?A4x&U^d7
zxR)`mR6T{opi{fRm;8KywZQ#M;o$mX7DXEpPK<WNG~{jF{DX~iKDXjn9yYJ_!OepT
zJvKs;yWKdTtf0_c@g&!QkRX*%1J(sWu26?;mnK>*Mn|7K!JCooM3L|{0J)~fkzvkr
zxZ8nqv(uVGS?qa)7C{GSOXw7b(bu{yk;=44U}8#NtDzQa+Ay_mR}7`bJdV_~iByG(
zgiWs*BdEcI@fHU>d<`OyCd4Opp>q?NDT2Y<V@I=BNl)4W>C7{q*LSn$(9s=y(^P$O
zrg(YnWX-8})NRiGvp%QbxOjlSmDOQ8&dZV!w^5_AW(&SyKIx4ddaLe|m!Hw(&Z)nT
zVkp8Z<2cP9I+l3cmDTK=kVg7IW9gvWxV~{=GGG8xPWGThjUOoBYA8Lt&%*WOw?t(Q
zpMq3H$R~DoX&Gyt1j|GcE6pB2)Rj!n@G1?lA`}?UTHr2xl}bdEEMLnkr0h*(Pn>m-
z6-7vWDXW?slHVV4GxeQ*{q~#4M#2PAuQXU}hn+zUz|I)%w1Ml!N?RZ8P1^-JBD7>a
z`q7x|F{*$cIS_h4oa;i8=mab_0{VqwcUO+4mM*<IzSotUXJ_0X;bw=k7oq_OcsYFJ
z$*+)n#hyZ^VE&+re*amFwwXJL-GfTcUK_({d{B-C4KQz4gRsmzZEGKpmyITvzyt8k
zF%1+lrO#Izn68&wzk$nco!L9@-DOctwAE~R^)ploH!*6sn4>_55Z{E@-<`b}m3gJa
z>!{j6+`u(-3z@SLUNEfRA0zqxtEQ`eIT0fviO!sEZ@Tca$gG6~uNH^}{EV`ejylsn
z9PQr{oM9^?J1+cP)k@2#+J~mcmIxM8Ze^3T;1IAJ3==eont>1d@4>xbc6m1@LDY<y
zCmQHn6f^D6TQvWYNVnL3o6-ZNOj3M9|8YB?)%>?9y+;2S0c_C;L<yjc5Z8y`9N_es
z)qLE}#g}<j<}fd{jsCaV=zkF}+zaoi-YgNkTY2~Vd%}wm-Fs&6N@s+)J;qq9M^nln
z-07cx&`a(nP3-$3-5=SL(Pv7BeVv?qXE#uN3fOlLMtBPzC=1%X7`NrEK!edywu1T6
z<XOnpmn6~RL4H2`JY1##iT_YRv(B^&CP|CLFOXm({p29Qcf=Xl;Px*qVD&Z^!b8)8
z?{m&!(?z5DJ@78UuIM7_j~S)#O5>;P<)R0U?{{cOO<;2Y0CQ}*-FZaxq*@`YF;tk2
z$1}G&uTn?h`U$7h_zxD>QYitFzt$LB2Cc^I+DZtTLd7`F>=Z`!?P9(X!yyb?2ZBy-
zTi%<^g4q)oDD<>fiZkHh%=(J6oOw97eX4a={>_-j_4w1}yp99+YkNM=lOj^E1!u_X
z;QdsuU-MGUZ&a2!AK4~oeeECnq#RwrxZ~a|>{Zw=-`9rP7|9)peAY2#-tRC8(CSVE
z`p?}v&{xu6DLWW6%1*#@jOH6o^%))U5+GSJY(}1{^&>{5Jj|BE!Tq$@2!P$){k38P
z<yQZ^cvgosc4Gk7k)#o<6MHXHtxgB^jM)F+uxj_<s`j|KO0>*O<f37f(V({w!0^Kr
z*)FEm9G71p%fxnryphBiua!US_4aC9lwZi{M_p-AMjhASK)f+<CDs8=E-IA}NJUnM
z9igz>-Rtf0ZcaD==GdtY!2Mded0N0QbXsnh{~WK`k|%ldrJJTqS|i2S<)*=mQ`=f4
za#$QW{+hCNHECdgq0VMs7eRr7T=-G1mdac~u^QPd;`52=SmEPzbyF@On)K?!W0l-7
za^#et*jSkp<Os5*(Exaw->=^m-tNH$QuoX`PV82X@w7pvxkPYe77QdG<|^}hn+0X%
z7?<!Vv05U_yQ^<HzBiRaDV3VG3m+{XPY&dXhjrps<s_(AbH{}s$pf@ZiZ;<XyJ(k+
zfXb=)D<`HcGGaI9)72{|QK@i?-sashc2_l&m-TY8z^vFRvhz_tcJBAx#6?<^0m+Tb
zt#~yDfAUDoKo(<#X#hGM)sqrio=jcjSJZ~^gN)U0s0dZnpLOWKASAy~sTia)YH_&F
z@qHf|RKsBe8)v$UUribF47VO3a~8!#@;w=8@%xbN#GO-rELAO<z;$|2sMr3?UVCw{
zwooOMpK&cV&`<yexdKpYk1K3$`NvuPnRl%(!>4qs$`a3f1ZmK4zOMJ<svyWJyh~}(
zf80)`G6`1bEk8FKz1+*DCK#m_!2Z+_1tSDrl=|U6hl{_rQ2rfJ`A66|kzIm72P*K5
zb&XJxK=&E~SY|oeMdFRW%p0(9<L})RE*3&PVy%A}DSB17k3SQz`~^l%%Ujr($@FKC
zZkoCS!J}LmH<}6l#CVARS{pnHih>_TE(qZIJ?M-B_vJ|Jv>7c!Lm+yK&)ef4C;vYd
zuG!<s%xEEO!e!6x*)_!~Im|JBl56lRNjVo-CT||*sebSk5#1M9?@x5t9-RU*Veta1
z-n(u;1X3i}mHB$OP=X$P8qu#W&<P2J5>Vb|h2p=KRaPEuZ~pkv^=uEX%4E!7UHIiU
zmKPZlI~RN0>RJnfqXk<rlDD@fs+@e_(SN#XBd`#Ynj6INT|Cu`-~HF%ukVNzmEWrS
z{z6-vZQzc96j-EE1~`xVi8^!(%AIldjg=L1b~R4?{;xO~7yG=Y)20t_@k!3>OM`>N
zKj)Coz+nGrOsw%{-*Bd~g75uMZ)`Y)G+|#Sc>=tHH7CB+Az)|oe#ky-rhC#c6q|YU
zW)VyjFJ#rQSn|UUZ_R?Hvv7OegXVCc%oDkv!WymdKU{5PN&eJ6v-EV;0l@EXes(Qk
zqZE*^_czH9dr&7QV@eN)*E?TKOtz&GIBiE-!1#}2r$UDvOvJwoIN0Nx1O6no9i~Ln
zWvBHRu*M@98+QTvW&P|7KpYP%L<=vgo_t=ms!^x%BG4|YVVg}XRCl4~c|3D1C7b{W
z10VV3$LyDxbAvba@2K32`Y#3eEq#Z;YKZ*tEg1jnj=SULWl8+WKRiGF0#cbbr!!-@
z@2x>FN;Q#>nXXB%(rn&-6(I5Ji=-nr7keuSi99Ep#*8SbsJN>fuHEqd5|D};$7im2
zrg$+=IX!bZnb>W5_DQeaDy=VYOKnjHRxSP!FYlb|Z1R2|h(csoj_!1`7n+{-3*fdd
zYV@9P9}T7^T%9CqRPEnnC%J51G<pWDH3y4~T%XM}7v1ibH4enGoc%nA0@_X+KCcU8
zg!SQK#Cjxq{N9LkBWYDYiMHpVZjh-#$+1`!EMJ-Anb%7DijDkr8vvOoS#Em6BK#Vx
zrM%S`Ez=}uuAKMiQJJms-RrPgU(;rQO#|8*Y#QEP3~1b>`a&cNO1WNR$%+Sb#X0f3
z`qk<bI>^Y)eaREMwF=67H11C}8E!5A*f3HvhxDBF<}<RM!XqEv>9;gIjTX=TiUSGC
zHx77gmW4!AwNT}4CeYB(5FgI==Ne-f-J5;h5_w!M?BMzKBqVELqM->*cLAtFkOS`7
zJ3YmvuSrRW2H(WU!RZ)JMr5#Koeh<OhMt}@G4S7lV71nv3A5VVsMxTkp<7>6FTihr
zZ3Bp@yX#ZPAn2bqTX7otSoS4k<iZz02yjo)RQI_PT$j!GFkRk4bL=5UEuu7<`4>p(
z7_VtP=Y4^P<#MjeG-^5Nkpm9DVjnczB%G|gz7K3b0MDpuhJco1Uk$`^0j~1X)MifK
z)Nw42RYtkiA^T*z=1)71I2|cY>XU~9Eipe-F$@ej-Hb000TIy~_1x5wl6*^7_j;F!
zA}I5N97R(|sJmF^R{AP4m0uD0`KE^bZePJ+8S*!@`ziYo+TAX0)D+{wBFY@GNH4CL
z4|YipRCrby_ekvy^3@dN6n#WwvqxGY@%G?LlDVHN_#uhpR>vFb(~bpi>OG~a&%C(E
zzl|I0(V*E%9xXwS$7A=2?$Bu#Xr~3g590)PcXhP+JMl!=>c;@H@FD14C6B@D)M)Pt
z@2Aq!)!l9T0%JDa(JE|I(eHS?6u5D|h0&rgJ~p<XKwn>9Z(~zLF-uFEujRBPnN=_W
zq@!6#z#wc`uLQU{y^Aev9_(s#z&LbHVMmS*{s1_YQYahhUELZ5m#)~@=HK)Gj|Rb~
zYXgm1YHGy>W!xBo+D}k4#s*!wD>MsJ%_|-GG-sDb`;-i&aAEszuJ-~Z3{LPo-C|~g
zM@L8BG`Mb$TIVZS8t2`v$D;r$`J(OlLX{Cd0Rb?gosU;lsQAxDY5myB45b4OMMo~l
zF8Ly(Gy!@{3lX>;Z|R{4*wnN%xf0g*hEQ4`2Sq=(Kony_y4au?D*pz2jTRWsKqOYM
ziK$Y_NWlp;2kXTb)kK~1)g>!^L5ObPUxEq2PK`1>2Ot@(FPEy9pai)Xo<~+q7nEe6
z-w(V`A6~?UfOFfcfs0Q|khJ2&geNyo?+AV8-}3z@<l!ecB-|!C1cV<lWoA>;^YyOv
z7xShm0b#3G3lH<5p$G1}lhj%D%5(uy>7pp^dlegv_t#LD>~$yo9lO%3=F^uzwJZjz
zB_^jRt&#;9{IEY@Ez8lDWBTBV?6Hf&2q>9(2502gH&(C~vsvR3`d?Xxo~uM_qX>=T
zA)+?(xZUDoHfa4g`*A}FDlBnce$QqoUt|A;73dFrdo=~uv^5{;jDJu5BWEpiquDJ$
zpyl6$iY2*Tlt6@<a`a8;Y6u}0agC)}q1fK~qSt=QU}|m&qrR1Ndr0EN!UMf5O?^iw
z(YJ5kU?a87JZ_IBrlyqO46d%I`L9-AWihQf&%2#KF}_pdwfSPHHxjf+;knrKt?~X$
z>+1}lxM1BVwVc<8dc78;;9SJEBDX<2la^<5a0}^g*`RoCtrW!pUiKTJ2L}p>rUvsP
zQv7c+PxftR8hgF{o~@27JlNC-(?QPZjqqope`|JXiPBd1<%}ePQS;`HQ7Yq2Yunp+
ziS_<I!k3){-oJ+``+VT3?(1w9;+AKfL&Hoy#mAtmN3(LQ^(`Nk1OPBNQd-cOt;uDM
zyC0v?I3_%Rkf_Z-S@+@N9Qg)_98Rqke6^4n<99N&Kq1<LBw8G|pEP1=dk+4mGL#1F
zZ$K8DOUg54iO)bDrzC*kDD{=gFmF?hOD%GI-%C{2rv%7AQlZ=v46B3yxt;F}vZ{1*
zYiK_I8o!c$k6&|!l@!ohNnR+(4&-*R7R+6opzd0C(Kp)|>XA(5M*6DpkM@>FIp}{X
zd|6~Xn-THTQWr8Z31en%qE)B)j&WhcYLbU5Mzj5;KazkpxBk&~H=E>>B-m2V%A!Av
zyyDp2!L^G-nidP6K|w=B;Hg{no<A0d9W54<Z=g1OoH|dy`z=dv3FVn|mwY?C`-MF4
z(~Nq2tgo|uKVk2Z)$+@LxBg@grhR3P2gx+jaA{(T<L+IC(t)CUV=9XMB`YSXdPg}M
z1%+aP9BR6pUS8Q4_swDdv|d<^ww^LR(Kg{QDZd@OI1{$2o2z$V8&IdO(vV;CgTAI&
zB~+cWtdo+i5d}|jk8!u<XPKr0yyt7VhdH)V@mo9}-45U*RSh$L_Q=m3ymwd5&!sBl
zdNT{<H!M!=zMFfrLW%P1ncU}ANaYkpVsb9#?gJZ-w;Jwjp|qdi-*p!ANTNbp<9W->
z3)Z+D&L6oER;6V4+|kRoXI&tom@sLebWZWE1n3c(*Nb0-AiD4Wm`WdpYhvbkmFrj>
zb`mXSG=sDe?{>Yk6=W@IBKgr;1uJ0(B<sGs^|}2;D<WUzglp@kmFyR`I&#Gw<m#n3
zO;BlwJ%{m}$m33t-?O;HYT-@n#`c)LwT|&wt!i*8(7=`niKFLSh|xN5zK|Zc|JVi5
z`obw2;?4!b*~19-X<FVj5OLn)T}=q!=6d-qpT(`-u1u_ue9qTz(EGYyhCZyP{B$yA
zo`k~$6$1|vDJ6D3MGRh?+(JMWpD!8iY4n$m{b5N`+`7@)JCezaqi<;Fiwjbz=;f)g
z5r}FX-Mo^k+Dv-*oIM=J5pvv|`u%7>oPlDU6rR7Tbg0Bzg-Wm^HIHF~_oIVdY4beN
z3H}VBztPZCxmoGn>n&K5EL%ycS7xOkRoP#%PG4zk*cAMnwx7tHMof}bB*?;9gYp|8
zdbqQ$m$G#E+nKIE*EYHylW>5KYv(p|83b#91cTqR_xT%#V0Y*8G5#`P(WgvvZB>bx
zzLe%6UuT-W$ipzbWmq7U9R{IJs70iY|4R!XYd@?3-}dGbHXIQ306C?J-109<798=$
z(lrseo~I4(ZDLP1?0=S(N}~`Zj8;iwT6d?tsMF5!Rwe^E)EJwT#?pc`10XxFntN`}
zBlLw5{hb3o`ry(Z&%V8VCW}d>YNOj*fAQ5%z|g>9KRDz8%#Vk)A0K|aat*=faiBP`
zp;nGwU|2x>(F1cr(Vg)<d14>Z9lGt}NU=$$SrhTEv)B7*)J3fMltQw7$fRms*4f5A
zk-;Z%x#d_8nrIn`@RzqSov4gdL9_!>2`8Ouvb1P^h3X>ya6%m_bvAgdw4SPEEe5uH
z$VuX|ATN^7zBj-?7kEi3{OpEOxXL<Aq~3aOytk)K$wR#csaGznElUzex*SHhD1_Wf
zE|~@oOS76w&64q1(l6MJPZt}kziH$7^v-QAf&gk0W?`<9?-nBE@a8G@&X^Rk!&-Y{
zU^mS+=rt)HuH<7<dgIYEzF%uq1C1L_-^{id_+}O^#Mb^c%+=h&D^w1mxO{IL<6-Fz
z4dDthIvC^)a4P$&8$+YJ%zjKIYIIrzro_nWy1CaaR+O&iNo#22Z_m+|2m5-Zc3|2C
z!-iEp(K^VI1W*~7?hr_DFHX>|*l28&KCFz8OZ45bwBK8-S7&{Qy1e3id*xWx&OUbO
zq!no#TqJH7lW#(G^c_xdd<~<Dn!4!t^n1nJ6U}-(@5nu(A7qOsz1XxYX=Qm?glR8P
zy&<A!b$<XLwiyT@E)|MYsH}xTnCO}xkou3f<~5lq5VtbJOv|MpAX`LS`XHVoc+QEh
zlUU?IbN|pRL^%Aw%;ryyg&2Dpudg;;9q{z3&3Y)6IGuybV@GLzQR2+){%(4`C6d@f
ztTy5JT8Kf#YsRJv0>$~DXyrZBG@_Oel+k)2JWVkz@gB4u+Ud7tt6k*4A-!*|?^1?A
zcnNxI3UtqW0kS4eGLSVnhX6-=<Rq0?p=|sKr9UACqMnw5nUjl|Y5)taKq;5B#iwZk
z8Bs>W50;gnRh$Gb0Ut$iH6fSOIKD7XWUPEk$ivvEgx`<4xjdF#XY(X<puxo*b8U~>
z7bY$3V?d_BG@L`-(2<&dP3Kg2V9Q&fo5#b`-lyim6(N-nh3z-90n0KqluvD?HEU&k
z6Vf+QAy`ksu9Fx{9c^;QUz<b%02khp8(IzWHbUf(u3%1om&Rw7eo{4hk<W6<-EMfZ
zJi?=1&>i<oX@mLJqb;gqKN%7iT9f77{L;!s3i|O1lWZzLj4^}~RI?Y{&xq}5`anO1
z705D17cW3bxsM!?91)=5a=La~4T?_I>b5jwNe=LPFjzaU;U?-F9q*arKZv+B>h65J
zkPi6bX9MPit*0==vl%_qx%JA0xu3qgh*NS4v%?I*pbw?Jyy_ruM}Jw2LX8`|p6oH%
zzk#@t9-x#X-ZO34t6FHc6=~AhQ<H`byx76p_5fnuD2E3Fe6yAv&t51(Yx(kggu`f(
z*DKZZOHa;RT97=ES4ptgwpH$wH$^%c(X75{$1<;kTaXs_Uf*6`^?k^5xapMS49uM|
zSHzXzdvjgl7a{1b>k=sFexEtKGW8tUFcAYjV+jJ`?ndIuYVHJVuKiDleptIYIgcCq
zuTjqaV8Q|AlfWo5DO1@)-r)6moT2$bDc^m<#leX?9;$u&xF18V9al0@>o+v=ytB-i
z*vwPthpT4$?y!toY~k3Q)q+eb<UOiJxS0w6-4$BFxX6sZ=If&@4~lrCe!00`AiPJA
z6@hUPb@_yq@rH4L+@~>EL{k3^6WPK^w{(PkyD}K`;(S4VIx%8g^LKO_%z2Xk$W#Gk
zP1*aum5TnTvTM@W)<FF4^db(|b!8>*e;CE3YCnzIpF-EX`L|a*ZIKoOj#zMy1~!HL
z^S+D#P+<38)3<~h^5FUu?f(^gTTLFFWD|NhkG41E3r-{el=@)|)_}PpYCKC)LBV^=
z`m^kHHZ-(UhQ`}omI1oM6ZuD<n#iJ!KK++3MF_6ly&Thx2J`UuOP>AZ4>R#f4QJV+
zF+M=flQ-7MFB)P@vFsfjZI$1k4=3~$BDp!A3DslX!n1U18kGP!B_O;s>oLn>I`${H
z5&{ZGU!M*5w+@3E2}hB57l451@tCTXeu>TJvOtzI6>h(LkW5Fy8-RczQl$0ZyIjR`
z-EiFf3y4@9Ut>AkuSr3l%k4=h_)IJ?_DblRgOYgl{e&o(Vqbw{g$5m!X@k9$&28r?
zKw^n{TQW@@@+yKQ);^KEf0>#hI8v=Q76Kbwl<V%5jIIYHhmWsop6h(DWNfw}X>5J-
zvcYf`6;vX1T;N8HAB<QkWdm33UxhVRhi5Zo-LwaQIt3xq{G1#}YkY2YfYY3xV<N(U
z2szNFaD&+yiixSIy}>t-kZJ-mw{@*<+ROsU(a==GXZ2MKH-y{jqpXpiQ)78;1o5Dh
z8~h9j7>IkTvEsldTRzGGQR>WG<_ovxfgK0mTV5VVU3pha>DT35ypFL$-Rn-ST}xV!
z(2#t}aWbTs$3Q<hGnoTCm9Wsg-?!XVi`AnJEoiNuse&U47|%DJ&0%9{5R7YTXwYJR
z0q5^MkTK$?p+k)ls7ik@EA06+_#~_j+?)bKO-olOA^8~#-eZn~I#sDq40H@4M>Ky1
zeG#r}%m|OYrH|^>nqx&gC_#*Am&-L;^4S74L>BYz8C&q${IHkKqJFUFA&3?_Rgk~&
zEeg!ikgoR77w`5(PF>!7!R6;CavC!!j$0K;|JJec_zes28bCogVxZ42A7)1mgkqy8
z8*?v?C{*`h^hH5deXkx~VUXNy+A#J=Sl#ETWXN&<bObI(zFePziG9Gw8YM@B!?OS-
zM%j2G5_&E=x=BF%f?~jG{fi-NLylNFC6xJ;O!Wsh>6FspD{N<1*D)}>3(`l&N`Nbj
zI{OY~xeU{?{*pJ4^ZyJb3tsv`3Xagc)6`(8LgUdVVt6fo4v4AtI4=8o)t&s)Iu~5R
zi4Jku5BWZ&Z9>Uh$({p~B$;bt6?yq^S2}H#X{BO5tm!2s(3(I`>g};bMgfHaAMydU
z02JU<p_2Nh5?`ASlDv{L0`&eCJ>)Z!<VHBk_GJ|MpZtEwZb5;{e=cE`_{#r4M#1>x
z$e9kTVfp4!P__%K`IdF?I~_P>F-=B-lu)ka;HwypppLf(#{r|`<4TH({O6N*K&Au%
zWFYd^fjIW#OMar_#{;(#ge!^tqcawUDgIR}FWAe%Dm812A9{mu-UGlPaH`?p#blMF
z+Q2%NKithgOjz@poJjZ;$bkv0`S*|mlcdp5{o0NmMhs=^cpdI`DY_7d`~6^3jxL;$
zoc8uTCDZi0yBpwtxalBjM7Zd}pzH^@(s$qour!&U{cab*8|5Ml-JrgBqEV`ByiE9K
zS?#W4*r&I!>02Ud&8w^$luN{WBF}0CL&OOS==)#p!TqWo8?!fHX$X8L-O8e?{FC)F
zL^-^=CLao&&_96WeXX<d)X3IY50kdpw}p1$!@~}qN~jruP^t<F(_LGrZmqDuRruFE
zRA5p5-%ZC|l>eBH4O0dFHXU>Qk^1=m1U33Uo7=djPaq)>v>^B!iTvNw$8X`*m?@a6
zjghURoT&BgFwzi)DgN3qFyGeSohJh^5%?Zek;v`zl%Y5Y(ORYMo>KX03)ds063yWH
zrN@EAe1Z)Z_B0qLF3xFtRv+9Q<}U;iL_}SB!yNOMUb#zFoFLB?nvvlLdM?0M_fg^?
zs_Mo8mfJ$(ec5hw@C}*nQu>9YPaLlb<^6pZy3efFAKL>Im4UpZg2H8cFuxCo8Vc4z
zfGq*Ej7j}tFhoZrieupePG8oZsC3doq;o><1fE2RdMCqC*rx*=4rre4`=k1Rzl3*R
z`Q=~{d8-tL$)ues!$_s&L)=pB=DN?H^B#AhkUnh_;H-4795q}n-}pZTDLEfI_(%O;
zsvgpC;*a3R4!Yh^!&Z7?(pmSO0m~kBBNzJie`pVTj)66-D{9H`TKy$4>S6P+AMS5%
zH%Dnlo8em+zm6V_AiVSK(ep=g^^I><?t0lkLp<k^cS6DlSV}_Bv3H5Eu;Im1lD~sE
z{SGq?Pgn~H7!IMfXle3qcTP}VJFeS6-PVCgslFs*Sx!gA-rRTI%3hshG8Eh{0fJ%m
z={^kD5fz-c=4FL*aR~keA4`u&VbP&?t?QSv*nl!52s(!XFP}$$P&LLQdy>+wS6O?|
zRvIbpNyP97QoK)KFZBKPmPpB_E-f#_7^=uo{svg+CEEODZE?2O%TaTpYWRJ6rHn8i
ze4}jcSilC9=pQO4m|!sVVdAH(r61Kx&T8y0k<}1tH3dqh{*SBfiekQc5~<VpS2I^?
zJ?wlep(q*9mCep5^vF3xlP;b#DcW-4%M|71W*g&mW)7Oy=;8gOqbY=@wF2dty5FpZ
zb6DRA3S_p4<yTzChBE`p)K2)&Hni_^&@=Ek5m#Y_z3u$Mn3O-H2QlX3Gi65d*a$Mp
zhTD{##LD$5Q_Hfjf6(bw!FkK)SC39)8|8-Jm9C)iP2dME_+LKzC!PGzTlWB}B$Ruv
z=cMhc3Rjt!J=+4`EKCcU`2J)l?vT-cbZWC!<bb9TD{2RBhpo!L*opH+9Hs&d+_ooa
zn486`+;`glaqUts25XnS%|tiA36xJF{^A0*gasmq{oZk5{r`Ra`Y+fDPKmq9<|RZY
zb%?h|R7t3sL9UbL<#Ps=Gr-=qo;*BChabKI=XJtMNEF?&>1Q-w-ddJz5;#;DsK1mt
zTC;_bD{AtE6UBB?eii&U5x}e+vlB;M$FV8(PSZ~CBPEXv8T%2i)0!)^K<cOR2p7OJ
zZ_g!uz!3~v2R%A$DUsK4JkzTQwoZ?QKY`hxwnFe~cd8BwF2rJ{GP;^1?wNskO<!90
zWh*XHP7M68@$ahy*F*S;r~zhaRBD%B<Ffc2m(^w3fOHO}LBPPg;K!yZB&6Gd^TyiM
ziJ_=S!23*eAkf*YhPAd%&d%Bf2fJ`!jY!1{+Bv`UCs+Z$69?YUrm8@Ea~Ry$#c##x
zxR5Ge$gX_aiqGILI%2XF83))#imI3`jKHxx_UI3zVHU8f4#b+q!x7#)z-C~eASYK}
zFaVYtSTa{eq#tl*?HtV|Bx{7Jd}Amsw*Ze(Wi3uma(doR`DG>{WZ3%H>>ji{_9E(;
zzej;RUJRS@{<YfmQFTl#;QkIHwK?jQ>!+E?Uaq9hk?jC=In2PGH6gHG&zNyCF%PGM
z`d@7_K0t)Dyxtgrqu|6p|5)?9YOjxi{>FO3JAuX7pMmdnaY=cpw8V?(<pdtJ65|OH
z1JSa!>70kNt|cDA3-I6=B=^p^Ze|n@pX9=2(`#0=NN+_(;TMl>0RXJ2c%g2m69iNT
zo&Gb>GMZ7{OnoGhPR%)3C6%es$eUj`jFW^H$p_L>FGEH3!Ksi-$tY|ATxx;IjIdkn
zc9*n04ZGg>Z|Cz+D~#SR&B=g;tbhPGzU6c1xMowOGwx6rHv4Y54Aw2x&z_z?N4%67
z#FOY%3nu2pu51~l-rP~Q5NnLV!aLV9NXGg8eUO75al7p$MDoF?dXO?I5k`Z{$$i0Y
zS?=9OaPaO80Zf{ET1za<DB`nTz*G1lck__hW{1!f7cfv4V<TQQi1`aNJs5)q$<MV@
z@&Taw&gj`oC&rCl>3qjSXLYL|OGbm}A?n5BuhgnU$yTLb#3Duxp)~`Km#C-j#(B@8
z8g{*Iu<LnmVv5P&FD8=OYY@qoe0U2QqZ#ndHLBIcYOz1l+J)fQaO&!&O<N!!%_ayX
zH5(ZFg@1Q|Of~{3`?`%))!{C^$~487?`}NC$_&KU+kBWmDMTrapYDw0w<(6PPa+zS
zo))!Mzso{@*8W;vlvzAYTnB}SUuhx5b-USEjG>JIuedg0ap&%^XWZxe=j4gKQf~Yo
zvZ>m#4QE54MYdz!8Q_{dmRBT!W&ccVXeH-NyYz+Jb%6H0=FfEw^8%ot!sO&0D6+uO
z#Q5UZFoEedHLdZSKAN@4`6qJrWyxr>R9H{QQo3D(>?zv>-aJ(`3xYa3h0m!T;DrBX
zJie;T8&zW2g-w_3Y;@5(45<Q_1Ej6h?x{%gFp^VVysUPTK)T$O(8DO8(MTT}n0PvS
z4rdsO^7I@kyyU8ZVq{E#K{+I%=NuxtK{3;ShSqdDRwyMIT~kZN$#W#+-eW#;tiKR<
zZZB3D`TMB4_o!4ny#~$U(UC<uIV*7@gwo=WcW<@VfN!WTlO?DWU9vq2``oNrrOc@B
zcWN|E87;=}n$#-ECRSwTXd(D3PL$b?!8?p}N5W1WSQ3PI90v3sGGi%^a#nN4L6oIG
z{>3F)?)xJSJMCuVv-1cheK7LL=O~#rAp>Hf?Nw3oEZ?pJ?rD#SOrc8IYhK=p;HW__
zCgHi(+<tH&Dn{7bLD6a~?^qP;XElM)G<|c%1uuMNv8Ok_N+&ZuOZ)tzUI>ix^jJ}4
zQ*UX|e4qEy4I-q%`oy4-+&^4d6HEq`QHEB#5MZ*I%E4Cmt^}Y!Z`G+Q^l!Y*RG%>N
zA^;|SrS<1JpJ4${{l-fw&ii`#BEk~jHFDi%$Qo<Yf)*7IU^XswcbWTVNvfyfr+I_4
z(`!4b8<D>Ay&fM}3kv5K)h}TZBFe*6BVk*)n0Zhwg#PFQO7Akh1EnBS<pZ_Q=_t&L
zGG0_9e#A3I#*Ns{J9y8hE>!W$cE!$sU7pP5%&FIo$65<|IvHX(&?26*3kE)Wig0;S
zhB9e(t`Xaj+CkV;&&Zg2@w71+N~dp|t3raNcArdX!1~%v@E~U}7K25=mLv=3T>{Ac
z%#NG&C9qpi=?yQzhb0YfKdYLkwVN!&_kNcH&y*|9f4!L>5pge+n(`K}e*W<O=E1s+
zh|_{5O9oAZQvuE+*rH+>v<5CAuY#ndQsM7npvTVTKW7;Zb2y=s*^gxpM+fAs)czs;
z1cI;4WVGH@G(N1^1Svz;PVrDbW7D4Unx8u*4m)1Ap+2!cV#^iQV~pb_ADX59M8<c$
z&6EEg#^N~R{YQ}y32BM1l>J14Nefj6gX4ui69=f$sx7+7rFW?n&)JtkiP18`!5L$Q
zSK=;;OCJnTS!^7TFCz=@4}Va{k4iG)Y-ikCQ}kh#Fjbk)RN3zkigYJ#ccxaW^VQYn
zMrtb|>Ola)e7FX*)~8;$bn`a(QdB*GXYf%&)Dav0xOl`1Ke;Iq+?Q6Sf~(OeuSYhV
z`<-A=YyxVQCjrF@g9aZHj%YD}I8mItK)l4u;+rCRblda@GeUU<8gvNd8CvHp^=D~o
zX98?7*o_tMJ=BU4l|m7?U{Vj@%{9YS&#@;+=d=cDf!qjG|CsN8djZ7%#zOoP<ML$j
zm|3)f;VEjmvJ?M(<lrRwp~Et>$7){2X~roF?PJGB6<r0_(BwN=!0y|vJH%}`Qt|Vq
z013RP3?MN!e}f^se8K^}WhVoiL*3kB72CFgdT@-L=e~Wyv2hao)kHakE+X&~mY@hP
zAK?o@E{upCJa8#3idYCJ%o<{kZ_$C%UpcAHgVpHo65r=ZB`BBE3KxxMl}X~X9z9$#
z4hphveSIZ7?+qmiJ{0j;R7u#f%2|V<_innumjd?CTRhbbskw?LeaYZPzfrH&7}dyW
z3rEM{_%H!2|JgS)LtPCw!hthp4kLxh9c;Ol3hD=_-@=~KIPaexb$*h`RUlJ(XI@^I
zSL<yD|Ns2Z0&Z@QX-ekPG17VOuexI)Lr3}w0on+3bZ5MgNanNy=NKRf7%Cb{6Am5J
zQg`y-SV$(?o=?QH{E;tJB*J9CdRnNM-=dSBucXE4nQGbS;ZR*Kqn?`L+k!w`1$87A
zp7l(91ElVlsHvB(*+IO%?x>vz#H<IjduvTUnsk#3;k2?6T%zhiZ#~<0GIU$%_no=m
z-q`gx{}u}_hG7m(qS%Y=H`A5_)jZ!7*M|2?Y^kKuMT*Cagmr(^3cmXO2`fo+F)kb#
z!r7{GPeA8*m8sB#{LRCC(4clt6mP4!nE`?Hs<HhWdg|YOhuT#$%{i3t!@gUe)B|8{
z`r>#%u9S6KGTxD>|ED@)r0%!FRM!r;OJ}u$%H8k{N5e#~?&rw~j=3e}mdTphQc;?@
zz|H#E#|8spFmAwe1-rrKC@7OqR2X}+cG!H!3`>pxT?Z?<0;>9$4d`Bk9@V#S;^*ZQ
z*nkTs`!-zgSyKjLbf_0zEvRdNq+=psZP#!P2eVo%!ID{a|AytfJ+w(3W@+I1*NgKA
zH9A%*GpYSH(5Q&IP4rh}wfI4r1j}~WDP&&z@c5k1emg>d_yv;->lPJe8%z9Bvdve5
z_#p>zBXGNu#RZZ?XU$1wLo3n>Xb&gCf4W=?6{)04hPgFVK2r>)7?gp;oFR64mHM)}
zJcs#m*(4j3od0Y;hK>kEeS!gooaI7vy@9yFyRd`aE+ze!gIX$zvSlFO8CS_~cL2RS
z${Y%GF3_QUWqgJ_I1bLdf@=U`0oSYw?2elce1<<t<}t7u(I;yck-}v&nj4o8p;cr2
zPo(d6mT2aWy0Gwr|L!p2p}%+(s_y<MENZ+b?k}%sLBqve-`e`1rIvmw3kwU!O{qdj
zP)^?*>iD;V6PLcOU3?~*l-P98Z_3MC+FM<mE4Eh=#Mq6i>_N4+fHmi1Klwb4A9J(7
zmtf?c*O40=D`zdA&9$M{pGS#`LSKL~G22ws>157u&)ZXSDysT5-|&@KUoWI;2U7}6
zFo^F-OZT08a2X%ZU%1{BT`rXq@6J#E1(llc;<^2<!aOKqkWpka%0x#ibJbIp0ZC;*
z8Z)TV8=t+r`4cU+lY)QtD=q@n9n{485uz)~Kpi7J#e9)pnN_(Y0-<uq|E8s-jj(ih
zd3my0nNMrgMsyC98Tn?8TacbR;BodosnoKsUuroH`wRow(B{J=lOZ}{QS<CnUPy7;
zqH#L9Jop;GiCRPqAh|fo5(!kz`Ff#P8+mdgjCUmRin+z@i1xv-oC;pW98PN}{e&nM
zeCbBdm;rV!(iRb~#*q?vsA|90IYAoz{`*b)uU`c|*3iX&6jjstarR-;l6^s5OCYDS
zn%u)|vky3=njceX;zlIPt$_pyT8&zT#g&yg5pg2{E~uiWbf{Qsh-5fRw2$+@;w47d
z!^BCt$%j{rtJH<sRUVy9pEkIF)V$`{fb;Y3Oy_aYLPbNG1eDM}gtDG|6FaH6r#L}n
zfoWgP2Pdh+XGq647FUW^!+?a6v?2c)Oc7ocowg;sNXGm+cDa;hv+3!&<n^T|ZM-N;
z{1e~J&8`!;kOT`&go$&|hfZwA=(6dBKq`MiAI+HOo_<KY>7S@>A4E-v5;44$;5hp{
z&<p|TTAwXvykIIJDQOzKhQ%ait+w3$a|h`~<WhJ`hD%gbC&oBB8x#bExIRu|gsrV|
zZU+#Gg}7!mzm@qW13x6f_ySZ-s}+<h8pgb4f~r-K(6h4BLk)j`7;bh?tbg(ZTIjv7
zpu+bXXc#Imf{+G~Y19~y2#v#>%D%+zi{(N;MV5HSH$XkBZFcQKD|f9?fq3$z=o!un
zl%Kxq_e(=600?J13+uV}XjA<5j?r<tQ<0(cuWCp_xgih`?~hyk8cVOvh(WJ5W8?+R
zr6UwB6+Hq9|J!ZJJ=K<pzY1_IP+TSxC~cLf50?kgM!aD?Z1N@zMAl!(nQCxX8Eon!
zLxEDt@B%}K+z!5P8%Up?o<|6NC2x{8h7SBSZ87yW?B`bpwdgP;t@^8`Q-l_|>oeHG
zsnQK9<!G6QJD|MG@vnhc4%EiAkjEBqQe)27uqvc^tWaXYObR_fk&y&>Oip?Lg*0xW
z$jB)GUw0ike#}0F3U>j4zt<(M_KB(#_5su!yPV*3cC~tI+`83h%5lIqyvzKxXJRjw
z-&Ng4_Zh}48NKyK#GOh5-utEqL&78o{lOzFyafKpu3kRg7W*ly&j79>en5`wuw}?P
z4~4=z@6f8OxRe$wPGEuyB)GNyGWG2Fdu#lnp|=?yId>I91jrQpVn#CodHg|4%rjMH
zFfCZq(&^#uq535{J_nY34m>G-9#o@9mFm_o?O0fdF9%VPYBmR}zc6YBWlNB>r$}+V
z&@{?qq_$mo4)jJX^)Ad~SMC`%GRPD{@=xnoOue|>!G<Dho;M?Wq5otAMJSM<#KMV(
zes^a1f?wOxXSvD<p`-DBLJaC!q`e*5Iw4iMwyq9Zju*Qu@{RQ|gHjTw#yhUh@0Qbx
z`^(}K7Ad6TSPfP?Pg{izVRv_@PHxT#rfco1PH^60PDs|Q2JyCa&r{r@k@?u6%Q;(h
z1m<c1c`BE#P`Fy)zGy;hcv{=W_XpO%5Kf{QIiIsFu_%Q2m+qa4VSE2V&dDqPNJzd8
zagsF~XOdJjBpQNSQrtv{?I66EXZCUroZ;(;lQAH%KT~UuV(Vy~M+$nTukb^DW|2=u
zLaICg2}e76&~`PI688Q~WaKQ)|L~HqRdD40<`_zB0`6mb6V~7MYUevrTnL}&N>N}X
zyHC?m1|MWB|Jx%F-fR)-T6E(nAy|IFo<FvkV%>@Y4uY3Ys%8RC5F#%Rfg@`RKsb;n
z_#YkL|A}S$N1h<UcAFOYGX*eVkd~|x<lYY{Vni<kp&NA-RtA$O<2&>A2TW;TItzy3
z2^<NFxby0>nVPf5QHBAZ6LpnKfY7{r%4Ga?dX$T#e~Z+KKO$(-Y~}>N?-Q*i!gu&z
zMpY_jsvCX*N|eb)qY6N4`ZlW8bdf=o@C}5Fl8_&zBm_Rr$w-O52dF1c9v=D`EBKx~
gc>?>mnm|1<8Q7rg<#%TV|MKMhTPfipA#I=k3$w>CiU0rr

literal 0
HcmV?d00001

diff --git a/docs/assets/light/drivers/RmtLedDriver.png b/docs/assets/light/drivers/RmtLedDriver.png
index 62d3058e93455064aa9d3896503be29705e2be23..211820a67095661211f9ff69f41642c40a9f1f68 100644
GIT binary patch
literal 27347
zcmbT7WmKDA*XD63?!~1Pmr~r_-8D#YcXvYZA_amJcPQ@e?(XjH?vwxftY_vOT{H6`
zAI_a*-8o6l-q-%^YlkZ-NTMJTAVNSuph!!JsX#zLHiI9g2vFcFm~YKA5D@4P(qh7D
z9%-lP@BmfKmjOjPH>gkE!`Ltw!s5M<+qiY-wq<km^{%yb_75v%_A4EF^=_Iiwe$Q|
zYr2k+(Bk61+uRXMHoksPH_6bB2*m>VHYVP7ot>P{n>cq~&si1Dw;CT&zM6SJ6$_b6
zY$4J_h@!&?^KigH%b(0znWktknp>UI<wniMi&vr4>g1Pwm4FsThdJ5|pi%)ae`(?)
z4CX_lqjA!tr`6FdUa(f8$AA%*!$Tnp=ZQ))a_&_@E6Vcl`NWr9rw#*tUomuYoS~VM
zFdeKBJaid30+37!8t;}Mdod8W8oi!H@vjTWJh3v-v}n;+GBCoiA>Rp5$UxZOHw!aB
z>|qL{_kK}g0^fuI>R*e?k7$Hx7%hf~9vtUJk%%!$`GKs3DIT0MRXoD63W3Jl+J_ta
zx7=A6@XHxs`>72Mvlgsp|FYx;Qpv-^aUt6uib3<UG0TvyDlfl0HM}cZb-CR~%v#cY
z0+KxhjnsT6xGT7tK5cTB2okO^ty;tmg%8#ckntA$m=~0hm-7$!iU8Dt!=u=2$0*6R
z#0jPNTIQE<@AqEjpRf#1`OkB`gMSVl=Ad{izpmbIMo+I#blk<LASH|>ipl)cThedm
zI2UQpT9#xc(C)>y`o3=4o1;eHs-ym+T0m7nXY)FYOzI@SX~}4jP~0)DJ?Q1(=J2?C
zzw(y^x*Ww67Jq)I<`Nvok07;bEry)Js3*=y)HH7QqHQ&9BpfW^Wq7<6o9|u@JOgeg
zCku1&)|+^AVO8B0rECOsa03qCZIqb#%b2@mCp#NT4klv-)|!N2EXAME=SH5tphl`c
zP!|16daj(!vlUfNXq!!}b*715>-=&S<z}PT??QNBcTsaO&C^s}D$QH1o%SIIAXiN&
zU;ZGm<mN&0g(dni6ot5)#p#6dQ=F(<JdP||Rg>34eIjSCK`lRA#^@^pabNc0WMgFK
ziP#*@HH3qE<xRq}*WwCRZo61ELy216?&l-vP^yT0Kk5q|P3_<F6$=eH3bk7ozvx19
zaOs?J3k(Qs4ytEaU(*5=|3nt%cEp-HQ-9ETLbzN#5C}R6rIb2-RNZht_2m&8QQ-1@
zH~y8|1~dF5ZrWkDS~9;XNWUgw0R%9m-|@zHKbV^^<L&;f?g`gr$|I_84JDh<ha_36
zD99XVk|sB6Bb<4u(&bL_=YA!(<WbfO;g(*Vpt^u=D)qFh&0wOLXpY16zuqtRa<9c%
z=3Z`I3Qv=c?9TUyfU?CY+vBZiqgq^X*{-22{~}*l&pMRA5r2kfWdGSl6_h))u3Sz#
zU&&RVCnbwDwd5@^aY<swq>xemC0nRuO35f&CV$DD{U|a$REBQ(B=5VqWJ|SE)B0tM
zHU*~cZw*;puJvwN)|SM5Q3OxrXVb)rD|$!jpe}4BI9c=GU%Y?x<K*h2p>#h@JEM*@
zJj1&YqQ#(zFmV&zUk0{StT!4{*Aa8mIK4$v$f(OMck>z5=Kd0giyt@=!5>CpAhz^`
zb~vU85~iOpXg{*}c6NI5-Aih=zU?Y6-?6gh>~zPI{5$@z9aFUaxxzx+OxeO7m2Oh_
zQ6W1|erS*Zb>VpYi(pAEyXLf_PNRLrc-5BQ+nf$1MG`~zu8__tz4Umy7-6m5y#NVn
zdwSYyTN?C2rj~L;ik$ZK)%-#2Gj2O=wSI-B`m|TD@;7EKIGOa8Wxdy}agbuQu^>{2
zW=zMQ)g7FQFmtGg;A&lv_%SBG?kG@C3aQ&D>U@6htTFpsAj2w3XY;^v(xS<RBcC&Z
zZF$MDhcAuDK(DSnPe$3gqwCEZ^>{-xfBto2!GCteu?%h#nJGn(@_ltq^y;&4rM`z%
zgo331^IQMm<z{1-@3}8Qh;FfV>N}4a8$q~fYqeo>l$P3C!u&~6nHF^L`j&8R=xy)=
z<Q_#K6RDta^*|fH9!V-++yLuafb7U#$;Tqf^3`K4m%qKcnS1P1t~pa?d#u{pmZDp_
ztIl&>9B)E~AJtCvdXf+IG0YtjHuoiVJ2El(NIKN~qAQhruLPN|@9;yWoj0nV&K$_7
zefN%bgK?w2QAS-s2kOs!kRB~&IHK%u9go*?V^@dmviZ}1w?RJw+f3hbl8F(Gzggj8
z*6xK8`^K<cSWcW%Wh7dL)aFg&hp%6KNhGwTrB?`?LRh4RwaYI6P(y&F^`kQIH>vL(
zZ&fta$lm$S4ZH*Q{m#$AFJGYGf&kGcH!#@S6(4iF4wYhuWU2|TE)Gk1?-S2WF8Oyn
zmDi|%b!PWLLOnk3&S=qQMUqL^+i8>=0zT>pUgOOh$nz}#M}<&g+u5jlQ@&*{UyQ%k
z=#3t)cqOfZ(Q|P#uO06Tw6k5{`Icg|2a>|Y>KU{p>_VpZE*5q8%2i%EbW7CSQQb$w
z)6(aDUzK9I55J9;3OPV=K0B$6SH#iQm}SXU7OA}+43%u#fl^7yF>yF8wB78<t;F!*
z_rX9s?S_Se7i72c!HV6c%B|05yep^~BKf|JgT}FJcQHn7(R08KSF+Kyr?X_g^Lw{i
znX%7Ezk>H9BE<3Kn>>PKY)#ZrKB2r_<SBowT^8!RJfLKgX7WsL3_!2KK6;#-)o;Nu
zaEOm(-XQ?)g~|+>MZX|fVO4C<GR5B=VPH=8ads|HQ1UlugmM^>vz2=n!_(HrFEZ~z
z{$3LXru1il5c8i6eY#^I+d_~JJ+G{sKqCC)OmHwRJ<$~9O!ry)Rd3x7B-{f4uLmM-
zb9&sn$U`0$j_cDt!Vqt`;*9Or?PVN+n*|wF+P3$lU-Vvrs@^n8`$!8;1LGadZj_}|
z@V5i^o;kIDvkEkz5Tjbf^qI3S;yXy+5^B+P-ATP|hNYN7^~ft`k8U@UFBZ*<tDFpk
zh(w#%5GWHWP8N9+THy@=aciLy-4rn+DZGrRh@7e{(y}wTx)Aa+bK4%|SGcKTB<x=i
zi4wL|@T-9Pg<i<};VfQlN1TipOE*oxAN&Q?+BT2eJd4e*=C7aI$Gg<&@E-?7yQx~1
zyhjr9Ymc7gg~EqRWC)Ht-v`}AeV++_J?BIK6eTyzcM`!tHp@@u1D5|4t}}2}M_f#Z
z!DAccE;cSRS;v7G@x6R)Q`t(O4THb1iq;ME7bvcLiETp6^kPW7f|#GaCX$WMCxUNd
zo?l;`?OZgcJ{BZ`3MiSv4~C0rzo%T-4c(9$<x;|~;~H!#>E+hKtfQgv^~X>Hr%Q-G
z!p4iiwM3$AJkqyf>6&82Y|n9s#>(7!N(H0i3)o4P^t&Mubn<TJG@1>N?fUEwaf=b0
z=X^SJ@tm`^6ewm2M2Ho3&nFhrOB<wpPxM1q2j#?Eevre{-q)J(eK%CyTot<hSI1{P
z)Ze-iHMI~`JWgweB2ICZVTY~jPZdW-gSi&=^#+n<e<^jw50%)h5B(CfcJj6rgDSv6
zA8<~xBZ^RL^rR-ti1_(|>|R-9uqu+s4Qmx<3G4>eMpg<#W{6O>&LWq=yaIp|tBbk~
z@1teD0^?nIdl0#pyCD+?%hI-u0MuRC&KzpSpTGN_+hu4>@PE9ts|b;p(3y4Z?C`Iq
zZ5s9aj4&c)PL!y;_qa1KaNs{;5@LX(gl>`{-rrh8(gP;+u%CtDDgL^i?$;G%0+q7^
zlpbA0rep^r4GeUe*n!+|PBTmFE(1%X!F=Dfe5mcu1=1Ruvfs)79`0b~&8qoUI<~J%
zDW1N#s7+o92={afoNaxEMwusi+LgHXjO9?Ol^ADj;mWVk<vje`ysvT`E>GRT(q?|Y
zA$sWK<s&yp!T*}8iWIkrVZ~HHV!cab<*JUQV{XVBzsYT<(1nScB-VOh3g1LZaxwZ5
z_rvDin^kH#3Jo+}4mB_yx9Mrexv)Fc$LLAwZ4a%dPRlvzm90>yMr&$nih3Wtw`YnO
zo-na5m9K&uW3Ber%<bIg(M{)0f8({8Sj0m#$D3o;Rutv=78%Z;j*<1-lY++B^QYXL
z5mQim`l_&K1)xLp&p1WTbKs8u{WJ<qOA9y9;NviF>t2O0wmVW&H7;%@+wUQ}CSC#!
zz-J$))eGNnvX~-fW*|kR=xKtkm9(BQZg>DB)+uqW{QUtz7~kVr$Gl|0!PiCgw#e?$
zW^7oqBLYC}xL#M8rM5qvUVUi@*C^a}hRGABrJ<9T9&vPM_hqQ(5P8&BnRebB7A}<#
zLnd9Oy}m6#g2UFQX;2nOPGKeDSm{!5qNXsvm~zg`5BxTz@I~YAJ9!e%vdLUi6IY(u
zJ_fx-HK0Evx`HXQx#_^^IQ+P<!HKEMIAh6j?wBnLq(t*~5VC34k@0PcIV2;@{${%O
zTecxJ$C67slb-F*d$O)cMhnW>fabt7;hK<VNXrmO3spVm)8+f-vGLVaBqw~EM8Ey3
z(=uqe>~n6mbr2-xb}tz=;*t$&(nt+RO3VmDZ5ZTqsq&*pbXFQoWOzpy+QN)YTZGY4
zub|NMlD-56P!3>vKti6J@fr@?Tj3T`r((s6N+~WDX%86K!>2EK_gsGXgupfnQHx2!
z+th;aNT%ig!SjQ;Q&FR2M!1o`bD>5@E}b~fxiIDnRZlFM-!nauS%BdC+{B8NE7CqD
z*XEo6RdSS}-u`D$`HGvv?PU9JejpR|+BhT45$>*a#IVTmfFb&&C;D*dea$0)$)}Rd
z76I|HDjrZyPf5f<iGkQP^CLC0TMAgHBu)4SpKyp){_GUjiGO6*jK$G2Wc++Ba_<TJ
z{UZ(CItF!S!`6MAY~mw;0L^ZxT4Q+GxTgqIZciY+IlOGU8|W}SsLTAa4aY^9jz;qL
zmsOjHF&-<or;<PmO9{8Futh17B=6f>(`mpTpSwke?GXc1-@)D>y93I3pCQAd0S&9;
zyFdPlTI3Ljy4ehcrPVYgj~p?!&-=*i?+O+6YSun_@(MB%%1*9P8k&j(Ri%FsqK)XQ
z-D_?_2BwFh!u&W~P@~*Uckyscmw#TIC#;qaO4F>@9Rh}{09+egA4sIHe#y%Aa@-lv
zAnwl**^q2O{gp_T-1kW%s!r8d<pA6WF_LjR{fq)NMP(fc(<TN^ilDJAJS{N|9k05|
zuv(Z{Lr7}kI7a7!=pyY7(}k;z29np^7L-kPw5aH!^3VYAHU(@(Rh^r%G%{oJNJv&%
zWU^b+Tr^!G#_ZmIKSnY)N2mB7rt<;FdOrcZcij5jt=)XC*Ce7e;AxEu_+JqgYy}@z
z_TM_}mgzU0UnG9c^1!4TF)En|ru#-2w<Dq0C<PQM4Hf*c)YF$2#F;bwDYmOsGp&D$
zIpsMSoFKXjS?TKD+7Rk|oj#f@u*TceFiI3tM%RUe!{I~$ONd=hqwBW}2h9Z&8xgom
z?g^2QFtiCgJ2o+>!L(Y{BPR2-$j4tdIbkWgghLn}WKK=`Hw);S&HWuHN`2@1oXa>7
zN7cRPivnpz@x5d*=0(1pW0&X+&;pDAspBM{B4nUfuX~!}D`299v2x4~g=Me;&g3wf
z6Z{dq+5oN2E*6io_|Qp0+~j*&6bF+Yf?l~@gLCvIe^+lz?R@ioSG>T-oF+6<jk(7X
zeviA|2`0yC)rJa^ZDe}@;Ijq&_`6V_$LcpCG)y<QMyr4(VrpvE83R2orLV3|_su<4
zW)Gy!@kLo%eebIJMN!q)W&3i)QX8GZK5r~)C522&ddM=n>mB+fB%RCK{rFdA10h~U
z7v#F(+t9B6gzn!`6}xp;BE)oG{+XKV6emWwdz4V$hJ|Eprz0I1u?Hp!IB^ngPfM*|
z(pajjPAQ+*3>SfYh+E2w`K{aM%n;_Q6C=bq-mwh8!hF*YgeE7g-KD>!11wnbH+Tpq
z6bC{@KDk`MIj%9LfbY+E9iP3Y<U4-qH9EsgCx26XbXrNdh-Z;UH7in_-F-*uJm&MZ
z7W9PZOnO;DOZ5(Q#R&g-&X@oQ*~l4nSny-PA<;TifoTT?Tx=bSkw6H(zAr>7Dk7oW
z^J4o~HBXQ91V8Dqf8$Y;XDuwgkk%rc@~l)lUF;Fum-&-FQ_wjdv}n4K^;_k4PxDSD
zE6A^${F}m~ck@Jy_ePmoBJX0O&?IZ_CEV9l;V<sgrLf_0^sOYD>%LouF;23o{6w$E
z_aAbrFR>QD5}wAh-h^HQ-7g*fMcbogL1amG_f($z{x4*F-BkoUp%TxK#(je<Kr9by
z1GhqpP~mU4TpiE%ReRyad~$AAS9_XV9Q;|;%-U|&{v7fri-7v6#W<%=*fopPT!0@M
z5+;y}WrZ~(NFy#z{_cv<6S0}8y7}z-4bBR28wd7Dv31CoHxM;dj!h<u_0;PNBgCar
z2u0u7e-(ArF?JN4PQ2a883qg%$dwoAPHa^g=X47_TA9f@Uq?h|%KtheT<_aJdgk-F
z{t~b8?7T7bz2}9|&j>T21pBYNKXJ*r>A8#QFpg6$xXS#d|MJ3!Kr@dOaXP?K!Qy!t
zlMY?qlng+Rz%zjDn?2d04ElCrUX|2*q|g1DF(Ap$?%KJH4{boUKh#lZf4=1G$mnU<
zSqQ1kzDVXKyo@7Gxi`|GZNM5ZQ!TWq%kV0R7s&F0;?PAkKV{OymW&l2d&ZsV)M!6>
zS@rvdAW630mjiP~{}XMfaWg*@%j{YMDYsuk+{jjw46k~_LA9sBxfd+U6#CWAvD?%}
zUz+!41tXbjpH{<J_;b|t*wCSGBTz~BDwpj33$Z!m`vz7D(#w6klH+VHKRym_i3ejo
zYwmK(@*|oqA&%)8=Vd!voTiJobQW=<=St)<wow!!tZ>UsUz`aI7^();ZRiG$Z^doz
z%6NPtz>~WmakF@$j3hT}Fent1)x7eM(9xYrmxObRaj5)lA|1RzeE^81JzhIG$^3}!
z7kR!50(H#Pwr9ib;eamuo{>~TxUe7B<nV8+_W1@v0!{>rKSnPUYD-}kLDb_^Kyqy@
z6@6l<wuA(z)LD(v@A9f0yq9FI3It~Qru(YenkZf~q3^}2owx`Rsm{DgS-<*|WX~(_
zZdLI$$xQFd7E{^M1t-cvQ2f?wX$|#cbdVZ#V(%qJ6e=rVO?tXTd4d9mOklU1rUj?j
zL<<^6mn~@N^FBm^`I2hCFx=eo^}Hd|qxJ<GqVGmpr5I)v5^;RK{ln)0{%?Dwi1j^w
z;;H$?5FNWM>uH0E!#vv4))1trsA@s`;^gQjvXtKA>6I?5nw_ssrOKw4-aIyW?QdBW
z&aW)FcnJt@b*Z}oc4x8Hn$*_%Zt1GOO1#$z{24||hvsa8SxA#W+_<RC+5NyR{Ca6C
z=yZ``rs~h8#BgO(+%+wwiFQtN_V&u?Gc=`ty_oPDlW=Mpr=x3T(U1RG{O8wosWtCi
zfNIO%xl@D3iN+zzx-5L#N#WLNyEcW2xg*^p#R8Oc(Y9T}5#L`t5!+S?Ji&4(P`YtU
zpKW&g`Rs@91}fxKMqJ^RA%YTmj8$zol5>JX;O&AEu)4ZrpeZ#8i=+J>ijfHq_RiOO
z=#boARq4Z@Fx!_X@k+a`>+Mi1DW~-VzV3c5dC&Ok@j8mLpJ_NB#OYxD-434>;L79&
zK*RH~X}4qpI+xR^4*0vG3`ek_q)G;ES=w2%;T;m$-pd=~dImTLUO?#>n()Vdaydhd
z?S`9+6}NT2P(aW7p~R&7%Spu**sf%Bc&fndfRL0-QF96TIq46P=jGTZR{P*eiyZpd
zBz?|2P>qhE+xKZJyjvdJzN&Ys^~rGI3hTFjb6Tjq>OKfp1_}^0il$VX<9M@y^-If|
zfn?}>F&7~FP8QOMOXQD6-TS|3ZJ+=FPz$k4wQ|XaX(DlMRjAU^)Fy*5wWpA9T>Zb{
z-}U@E)*Xatzd#}8P-?jN-s`(cD=0tGMHXfO!c6O{|H@^@I^yvqWABlW866<X9w0x;
z{{6XN(jCDBJ5s1g3>~HlAwl;4X<@_v`12R8L(}|PCTD=C>sE5&yhFAY^i$o(3+0ch
zoG08j^3wISlWwYhav!FOU<DH4Wp<4r(=7kO_utv`Y==sZ4vQB{D;|!e&!dImFg1Mz
zu2LZg9<pav-!KmbvIXj~bf*>DS{B56bKWPq`V)=V{jT|&POG}%>m0~{wHT!Y&BE{N
zZ9TwJ%gO9U`IFr1;TatZ-wYI)u}$E91^@Ky8lm8mlz=<%FtAiQip(~YYSqYCOHQmK
z3_T(TF>=IzrM1K0!Vcp{;N|Vv=Yvu}8>r7{EQ4irO2LR3ZyeC-Nh9z>%BA8D3RSQ8
z0d9f_zYDvYhv09i#;Xa6dzBS#RY|Tz=~^AHkE@$cxD5P@epQ89E9v>UNTx;+bxYdK
z`q>$bxY+o}8-wbSY|WWOw(cctD+vXz>{JntF$kWDi5d!+R~dB}%T>jxI?wsnf8z*7
z+>rO3%~uezP)4NI6CxxKJ?V0<6ixX@oJ!rR`3vR))1<XabFI!=aAcbM&nX^PkBt0^
zyAbq!zR3G)WVebE5^#G`lGlj6Qrp6a;GVQ5`so(nQS6T4AaP-zB$ur1=0vDac%2UR
zgy+xa=eB*NqC2e!GWXNzkfg&vJKorN-AkoC;#No8`$l-LL6=}0a=%l#WZgi*=Y3=*
zr{?ZmEMEq<-xqxGY%LPlDe*O+`)4B_9>CzEnt$uKY)bUmX%uN)ptRGRF-OY=H&Rqf
z31P<(H&I=0VVcbC(h4ZSdB1LznFUH>t7Yx;TGFWfE5qI)#$;5b9~)+MPi2BZedCm+
zxQHl$<Y_;IElZ!k+IyGm!GwL+sC8Kkc#4qwF4XBguA$}Z4%v4+eqeTj5;;<m_tR@w
zV)AjQMRgXl`0x#JFHNNhE%eWq?F&^)w-WZFT9V!dqB=}ZkE+>bm||9b#7EdZQ*dUS
zP#dOKS3q@9@cW*yV9QxHiQ=WHqk*wj*ixA4JY9DC2jy72^C0l)Q}Elsc#i}R<_LX)
zwwGuIfb^jM!A#X#ZP~M4r0NOxWX&_p6eq$csLDV47hY67wyQ#JB*tuJ3jXjxU>?ar
zvR&4PZGMDuk4~v_5n72nN_(-|I0ll2+WXG^luXbSz4Mqo;vAt0f(1uCQB9O*m9{P}
zeI$n*Lwn1(L)E&LDRaTQbD18D=U}B86IfR;>9_EIsdr!96!3(l^grbSSV!GVl65#z
zz+ZX-9QwnL#?x^b%PtQBA5(Th88bMQbLWV8efM0t-!zN;ENEB;An_N}JB@@Ln3LuQ
zuAAU+14ZC$Aw;i;i!CYx2C@lWn91Mb+$Mb6X95h>7Ye80G!<;I+GicyA4J-&+MoVl
zx4kFH;Ar(;EJrF9*BMSHx^MP*{6-bWQsL))$ke&-TSoCGajZ;IAi)|@(<s^QRh}xw
zZvwN{ftSn=Gv-&~*S~g#UnQyz%X8r+C{+jPY2Hplqv-us)6(L&os{x|(ix9BPt~kK
zkvu72TGM4v_<|jFg&^(iEKM_CuPgTc-ooXf=VYFrwru3Wx4eIw-O*2g1{|rBILmLS
z7aJpvd<JR2IdXqnL&DfhVB5_em?q`Mi>Y8p-m)<F=4EeD$!1OtB@+8eS#Kn@fPNFT
zIY)p}Rw&sk!Se1}2A6NjB503hF3F3=DQWPHv!E-d32Pt<kE}w4YF&!5V1>fng4Z-g
zhp=zwe$VmYuN*#M;FAh;K_+~vy1i1|)-N5-4T;U30zO(RW#0_0%b!u5F=O{xMdr4!
zp9o5DCr&1IY+Mr?tBXf;cv3m=Mh&!6)(;1Mv?w*|5MDDI7qe!h{lOu|secTYRw`F1
z;vnHi45pW+o%+3&G@5zc_jhbyf(jLXvPSbEYnvtIJ>yNKrfVxc3F@t_^CCo3m=qtH
z?frG`6GbjXcXNO+VsZ9$1Zs0EgL4YaEFsINoonDy+gio60`AR2^YRp&F;L-PS%?I=
zJ6f;lbc{)b=0e#@a`y8!OoeB9JR{d#gX`_b5Fuc2Nb20bsktcM`R<vR$xUgz;7-=m
zibkmW%~H3ENwej9aJA9x*zeGU0!G{ZtxEEg+<6OY^=UIfj9)UB)gU_8wYEd==&z4E
zGN6J)97_sZHQHI2Vl-V`WrC>Z7seVSjr{WVInTfzrfGNj1ameWW=$N%ucmx6q<&$4
z74j3w279FAlQ^K~a!<r2|5|s7l$L<l2_X<o`B9YU3A3laIy|^LU9_mMAe86qjS7_Y
zfe@}I1L%bXUghW$HMd=)V5$fbMnXwnQ&E7m2wZRAcJKetasBVwumAhSr2#(za8{xd
zr-}EmaoJO}NhuLzkXSrVD;)|C?#0G-WYgRi>DT0n_Rj*)yDZrHHy*7w@&%DYP}|xO
zR+i{h3X8$5RTmGy*Ypg$dvGPDR@P^{SsWv)6l-3=`y@?g3*Sb|81oGrsq$$4N;U+P
zaH?9KBvu^no+Fd`y6OL>jN&G8yQs|i*{fUI{vPCpa!E_WL|-A)B#CbQqlSwWj`6DU
z!Lmr0n)+iY$Ux_5q2wpqO#vfxZg;L1Y~)WMFsHgra>8`)kIN<qmKEk<KCI1^+Iz-8
z^5q^89}E<{Dk}Vzsp1<xMb)b@J@K7C{)V1rxhdtlNG@_<3T-nRdB37~c{nB!8=YkH
zW#sl9XpgTWXldb!Ccp%NG?O)zbnNYScHoG>@#Zw2kewd-jhCBuIlX<dKvVi;j+=pw
zA%T)6Uz%GumLq__rGtN^@QQ;3^&+z6_rF;{dsX&M4D}hA{J_}K4-2^pJ@|>dyaPER
z99G}eSjF}gGTcu#z1?k+<3nSzl*7AJGXom~dK${EkGqs+19W}97yHXHuhvzv6hDUO
zaW=}9hIX93bT=cP68gPqc*#5qo}Zvf9C~NIEziRG0AJg-cI-x+mUt`8JJtn6hWkQ@
zdyKg;V4fNRh0V4^WT8b#xifCfS&`+K$-5exVv~{gA{XrBNr>q;Z(8{QcV0c9lVW`m
zg%s{!opeiK2u<?=Yt9+Ek_4-$`bIU0(jageyy<3y3v1PBC^vB4onDzSoj%kmE-RAB
zUiW>JyG~~?(pmj2Xnu^~{u%xcM_P-|j?$|*hqyecAXjGtR_mV>)ytRpAg><5arN98
zT6L@Zh|W+vaEekhf(FeDplvxfk=R8mz3(p&%kC>a?XI9)7Q@1O#1OBOGKq0NiQ7c@
z?zruTJBLs9PptS}2I%uK@iNl;pCv+rNv`-$Sdqice@UOZ$PUeoFm#a2+0i*bh3dok
zL)vWWrbwS0+?<Dq$Ip1L_Y@SvJI$|mct>@^EK6MXf8`T{;Ig?d4exnvUE54bNJy)<
zwdb=We|EW`%1a~Sf<^$=5;M%mem6K~xT7hrbm{wJ#ZrxAXgmz<6&g5(u(Lcc0vapO
z*JXq5XXstkC|l7}QlqQFKn**^#hN45S#M0Qay(Erf?dxjXPyrCq=h$o<V>ZSX!(iM
zy)&~-x2q1;C8M-783J~d>Dv4I@}!h?I29X?yK3Ea$wkbUVHDj#JTG2$OTIn>?;Zx<
z8U*-T=8uP$x|rAo(YCP_#|U!gj^e>ITIDOc%l#df`_cZz3Togc&`h{=MuH&#8V5Hi
zsFoI_<E|;W)dt%N^~&dd8VTP5#->;~B?sR&U{pI~vR_Sm)@6d@CJuPGrDRP(4n|M{
z?z>r1i+zRhCx_(9pUc0BQZ}AVvZ{fbkC580&q-&WH0`5Erawm`@gQyNDH7h>t$62n
zYx?_>qD`w@M%pe4!w4@UmliN=ADv-Kl^TiwYEvWM=J~0!koWcdl$Xo}ZC?GxGI@yk
zO@z?GY=Cj+f^I*SK3gm3Ng*C>TU%3D9g}+i?6M4~!(K3)GZ0DU2+~11(tm#F0WXyq
za0hX<>d}XNZk`5!P_5sn)`1UbASGvBMnV_)Y_<Hny#S9uO5_VtCW?&nJCF1oNeNZ#
zJQuFZ-&Gf7zfmAJa|^$T24WZ)MVNX36U)}TAc&l~dRJsgO3ooiJ5y;>lny^np`beM
zJrjH84-W;Fzguj^)vej!WwaXdb`uGn&$Y>Viq91zEoA$=%57f;7RWRrGoQSH_dvoy
zg8*AlXBF(jbjU@qhX6!ZMNO3bGqU>~x~JJ~z50X;*~G)^QVhzbx7G0Tuf6v+(XO4>
z(@=RiV~<ax#RB}mlAUDS?Oh8UJW%_ASI|4qczDOLPKUS(NHs#4XCU-~{IF`}34$VW
zpq89r`3u%ga+J9ShVx??d}-PyM1~96Kbsrn?(Zg6!nOb6j-_+Co^Bd&o(C6#)a$5w
zd4MscGwSf)L3HwWvU}AnmRB;t5qcZb)mC?FI1R(YdkIfGGkqOcG>?Q6V}I?S>vuEZ
zx6PwT>pj-R!2Go<aORWrKd7Nl(EA#p5n^3xXtn;JF*`r;=Ny-&asv?;UY*!7vV?59
zBUuMq5F1h@y}VA-%VXUxTLD~-d6uc6CdQ2%9&s6bsC{qe;8;tv2vNLgH(RYVR^TrL
zWD|++oh_jjgoBhFcYD=&rND_Db;00X++L%RVfrv_Qv6iZub8v8rw*z%=Oc5v!zQ4}
zeFclJ{fN)0@dk@SYjxRhn|ElKx3P;6yC4~d(XI4&3Cq=*fDwXT6pe}v)f&vWi%pCG
zeqIwH#mlAm3zKfkkH;4SvORuL=q`pa`qO!pquiT?a?!tq0L5N4^a8I~4j}D{ai9^f
zI6*T*`VXqd0CPfT2NP47gY$Pv>Ij+PMjOPCm9tqJrX=XD+JxKF>j)Wy0pY!MiUVP}
zu7egY3J@DcthmxA<``s&T$wUZ@|tH<|8kv7DnruIlILYgG~xHn&a0n(804E(>k1=*
zCjoNh$y{{%4P!3=qBGKTHZYkhYy@!R`I5CRvv&^k4YgP1xrK~Ke08a;35GBYFeuC#
z7uf9MbeXLE#n%YK_=opM!Hc6;lu%F<L0iVAk8BBLY@jtIj@|W?3?a1}V<$xH$+O9~
z$%{U@AvSEo^-yj*YWC6&(m@Y68dsP)Q;~Y#1#(#IL<n-_uKV*f4-i76r<Po#;($es
zNp1@xd5=!ZyWKvfKz)&aH#U_n{MNJOC}cuoBRBTBq@IX3{_te)y<0V27?)hE{^HQH
z5(eA@=VU9;ytp)s-G?6dR+P%*+-HizVe~sQhSwt#{REnIFhP^0aIGwR$mtV#NLh=b
z(CGIsrAR+v!F}@qlV|_6MtByEZ~Uyp!wo7mWFcDQnxh#kuN@N;Apq7e%e8VLD+3QJ
z#P6Y>osnefIn%sC5}wgIE`B1%rPl2HR!t)BRJLbL^mOV0<s}>rp2w<u7|rebxBD%C
z6q-S(8|b}Cazk8y)*El|)cP|!c)DH-HVL5?AljRplk5e)8_yC>@FP}qAL)UcIkiMd
zDiz%l=y_wEW_n(Br%s&S7ZS_Z3Fa5_kLhekgI)`|A3YcTffvX^vvI}$<XJ)sfOvlQ
zVyUYfxpe|a7lHc+HGCKrLDx?@8afJ%9>k*#C35@gGzcSHUi4EP&}K96C!v)b9@GXt
zcA{2jz(gqXVK$ggoI)Xh1bEF~!EI#%#SS5&-FAPWzutDP)|<q_f+%lg2*M;(fU{QI
z=%?5TjM<mJ8ih|@4=3U{+<Bs35<NH!m~>Mtjp5yqu3?vewQ>?}<Q{KuqLH2h2tZ~f
zY~J#P(&X5xI=d>z>;Ch$B)^NnhD8fSC+Sg3%0Vg|(BQF79XJV+c|bQb2LS2|2^)py
z5jcQkgV1oj-`g&!Y+#BZ*X4=-f5qtXOotZ<6xby=w(sm=#1X294r7VOUzS~m*rspH
zRu5_s%>vYuWP$kFe`+Ms4bPx7*H`1x0BhxG%44a)qJXH5z>`o}E@5OI`c!68rj8+!
z48cjTGcTa4o0DM4NB55%);?@MlP=n2R~@4DLXK!GYOsPf6m)1U%4D+*Kbd~G02h$$
z6CR5#^<>OEC6FvkJPYsz><)w({*Q&>X|?TCg+cG&@1O*kxvAVjCV-R)ow4x?K61$8
zqq9o!+2fTX`--{g9tp2ovj#OEiTiMZWL%t`UYl1?ad;#l5%F4Kj!3)RT77G4_i^I;
zy8wk`|MFs(VtTW~vx^7?KN$khX-OlC#%drsGd+!njes$-?+-Ki`Qg-zefyLD<#cy&
z46D;pQ_V%qgn)*#d;_mLi`9^+Vb;cifxm?{eaQW<#kb3zfwdNRO%1xABJYn^6b4?F
zN8o>O776OrFg=nBk51u?d>YsB;}xrY>y>gF_?{w4Oe?6F{_ji{d!eM|t92*ne7*<i
zpS!*iaWL#pW-pIa2;PiZFH~&&ZHS+oBtsvFi(?VU9GFGMXMw1_KFH)O)qjJ}oY<(^
zDQsBx`f<8Au$2#sL^3ft>UlfWCy{&g@qW<(4Sx@tTP^$fLx-PC3JDWu&blw+;dnli
z-}_4N!!wf)f+nKRtR@@6Ekxgb&gFrsm#K{pkT{z8eC|);Z#y-bHLkCOg@wcMG4*}4
zL=W0WUz?d2@%L8Ogdn$PSK6nG4!#O^kIR!tbM=bOPkZNG?@GyhJojgmF>Oz;#>OFc
zlwm1gwl<O0G-yFt>2egY(lFEA)54yQw!OXW@_0G)2`giALl@6?GRu!mK1f4EJ9mnR
z$7OnIYQ3vggMX>obV1kNca*_9GtCFSor)wzvjHB)Qn<4Q(zI0llVClF-k%B$&R$Uy
zY!_Qx)guab)w0q<d~9u+k=Vycd~@AYQ>!Rv?MHh$EOAw$#a!3xN*~!l7`;cxoN%p8
zT5U{sq!A$Jt*{uKEP^nINw{5wFi^}6@TX_Dws5Ji6K-F5=(}l5lMK0|r%Tk6lA|If
zG%)z)(2qv?9!k#bejevVh(<6=`f6H`mXLt_?r|q88Jz|!Gur_IF;PRIGuXLydQUBS
zj2rBPH!Ui3>QIjh<p~dVcDUV69Uh4_)oG9K1IQSWXpBuv(E7KH7pqO(B_ud3xjg{o
ziqs3gPJM{W%tzVlB$DFo&oadrK(b(xfH(1vL6Z}2y84=;C&;4}?g7@>Y-c=Ij=Q2V
z&)nj3>_|p7WD_i<4)I#9uHkl6LIT^&_I|do-_{zG5>pV6r^ndJs(>=ee4NODxP7(V
z_aXwPgut@Zcwk+-hRSC7wr78=#mAR7umKfxf3LEyj)vA?o|%rKz=M}Tt(ZZLaJ>m>
zy4rLD@pB}laBq*L(J3CcZ)oZksqms{T^$h=rG-#MRi+_X+Nsh6!>ds?bv%}fz1nc2
zNr1(LetqmDvVEp3nv;cocu{vT$W0?<Bi_~w-=)#u7`k`8XWdKW+3HGv*B(U}t~U>R
zy*a`EFmGMK_=nZ}uSzkX#%yG@X{KDit;LUx$Zjy!@{-l-<?OS8*U^${AtoYX1OkxH
zP^0YG`0+B3gTttxRIUE9W3AJ9&canLt2nDDq2P`z$){v*dV0~+vG~`-3-A5K{%Pk{
z<(3W1118j_#X`%@9APm4wMEl{j8DHUMPgLO|8cj^UGOb<N#Oa|eR_+QjxHFP-J_SC
zUbEueWh%$BmxssI$=%g8<GXQJmT<gejNfR#1k45jU_JZ#q(U%2GaKbKMBwd~mxk|p
zkQqpZMv_tr3@<D!G!M*|)**P5MtTGcpN5po>PV!6BTp5o>>dk0FDWX&rP{B63DVV$
z1K_|wkP>cUg6JvBWJi3I)?|mfMH??r_|I=u1lsh`)i29w{8A&sk>^<Z$2Mr*{H)##
zx6<e$^A#KG`LYLlBOAyFmv;{jPgetqA|lOzMu&X65xceK+oPHIr7XvtwdEygvY~dH
zr4pkm3owY4j+bCna938Awy?PTY`_pq0ss&tG5NyMotSN5>_S3bfrEUQ@q8NIf(mw$
zI(ql@jiY;IjjLm|yR)NOpnGX9_z)idqE9#mt4qaSk%3ASt!-}uOaZ?=b~GfE)gDi!
zsZwb=ibf^WIs!f4ujQp=yC`m42G--jrBu!3)hDGyO}diYG~%?24ug}!LuMl}8@XS!
zZjk*{Fq9LS9e^5?9RPziOGh)D&l#=i{=y3R3>WRud<csFc=C6@!$}Jx)EJidpwQml
zf3pDkt+|ia1Ae6gc54lm(3kDVxw`H%IW&Z8G6;OnvpO^<r|ThT2)_yKrr`7PXoUi4
zvVNK1>a3z)S}6Z&uq|8@cc!WRon^P~9EGXD$brtI4fr<kqh?Mcb>Ix4s#i79@w$Bs
z7z90QC67dn=($S5FG9XutVCno_7x-}_T`74A2}FRgs0oP#IqekLhdvS&CFU`QhwjF
z`a|JKRw1vxh$`ZAO*}m<L{pfUts^fUO!QBJf9uq>0Vk(_XTpE-u$PXgWAHx>mOygo
ze^My8yM(5r+lR%A1@`@i&^zVSAx?92@?zGJc(o-`yKE9h@tA&MS`H$Mf_9^{pw!$g
zn$w7@ha5IN`N2l88EJk)X{6gwU!#KsPUxNR@GY1ja)jC@Y2aJ2T=vb>VuC-0lsB|-
z5_tRoH_$1Z48aJ4=6|tBxzY&k%UQG$2&L{17ekol8WsD}rXPeA&hzEJDF??tJU}+o
zKAh{`alYTV8Wi>v5#dftUIBWabcz#A!GjCu@X6TBT8|PGsrIrnw%x6ri>4=#0*x9!
z&Zg55rn4t$WeqoD>mhZL%HVtHr{E=HsUSrQ&S0LWn*Zm{ldN$Pw)ov4uEU)fU4nNg
zEalgABV3dU*#(G9;CA(OM7>jF^S!u-N5gmH5GStFn}dqe6?b>(AH}n7ua8^>r5YuQ
znF>YvcDH?>4<f~KPEYxQx*SgC(<47u@Q10VV!)6mj!<R}5QDAsI%XMVQ5LhYMWwCX
zt-xt7-RKxXhtpe7O_JSgDe~>b?x}_Qvj}EF%cD0xk0qWe+8=Lfe*2yA9yn21CQE)d
zkB|6+iaOpQ{f>8DQc}fq&#TAA#;%)vdH?I7q`!Z$(|KP6UQUhSScy>0hvJ?NCJgke
zcl9hdT#vPk^wq&iwa;kC1ZSpDP*;0N<6Iz$3=K=(t-t0bV*TZGA*8mN91N(ss7!|$
z5|6HS>=*p+d-|aVu%O=f29VEWbrQgy?C5*U>GH?Tpao(2j#npFo71n08l_r$dwW;Y
zvJ+~Y9{VW^aHwf)R<pTTAm8WPU%!y4_v_9YY?ffpjE$(uwQH8f6?h!CEr4XUprRuE
zLdA>H7T@=$mfbEPWd?9*Mwn`Ads;Pw72eqlVDTTO-mp`$uKmZ+GO!dD{C~lhfV>~~
zk<J?B>oTK3fNZuhO~3Q$C~(FeZ%@y9oo1cM&SZK|pJqXtw-v+#+jJa0NDSxpcRzoC
zX_NfPS$;j(3*}H5Z&7#qzH>lM2I1}L*(gVXMPG+Aw5v||&ZnXUc;CN?5q<+xD7yj^
zn|0T;S}Y@@XK<`E-RVysof%90Ktw%1-QHF!(H0lqo6e_qJJXpdRIEVyj9GHBMWNSV
zx01%Knx0Pl30SK>p5Eeef7;jXOJSvPeE0lcMM^9NO~8IYGCJ52xLQq%s7CNAKv-q?
z+}K5?m}M%HKfgAoOtZ;i`!7Ny=yR<SP(dxm2c9iRiOW#US9-7hVV|FKAJu#IwVn<n
z%_A;CVzGBFYK}%Ecri2NdK+7Wq#3tx!3CwyM}$Bi7Z&-QU}&MkMzPV$`Zl)Rf#8yM
zG@YZm6z~ovfpm3l>$l32r%CeOn^%)Vk?@LuuV*m!_KKI)Xe4~fHBL*_-!i<<%bKGk
zV@MEVE>kJ?lbt&J+Q9k+UU;<5x?|p=)_TE?3Sl|~(scHN^`^>n<owq!n`x)Z##NyX
z1w$pwcwXyk^M6sJ(v*Vpi=9u~b#5n+xl;(=#~PCOG~O^5eZh|C?8$FAsf8^wpcD69
z5FIsq^RtV55MK^qsHiK-EjX*ZIB!T7yf3Eo%1YBJCF9{}=qtM(ce~P%3dVE!7~0by
zJn}7KAil(;hejp4;;WgBobRANuf8sPXZ8kr66Lb8iW-Sd<);-8J{>{DaN6ZDD>m5b
zj_q7!WldR7DP~Nl`8^-;vdKq^iDWCqCTV(HZL1fo&&=3%0l6<zH?JC>o`Pea$nz|-
zF<~s_h49_a`IsHIY4$BWn{4L(y5@;HIwfGLTJ72`*T`f;QRBgdX_QcCO!_prz!Yi~
zLYJ9j(5d@Zzyg?BQjLq%iWY6$+FH;~2_@w0McZB`@g=?lMtE~0GP8<{bb*P_^NZNh
zsUTYK1d*W4Rckmw8EUn3Ay(X~o?P^Sb)=aFr?aag4O0o(Srji=i%e~7@G_3Mq=4ll
za^}Rz8ZS{^WwX4KlZ>?VQ)i%W*UvPbJIex0pG(`7d!P5G7$rBCT7!o3)NdUC9=}&U
z@L^3>FL}ok@X3mbN~Wa;F}~V!k|aUXbrq=!E7ShG7Giq($N}tQWB+HQV?$s+Vd8;j
z82jzR%$Fjn3jkV<(EJa!Gb%BNx+HA_cH@KFZzf=KpNcl@?mT$xW9D&dS)jN$#gKda
z5Bt?*wujNRhToTcF;COI4Y=#$xswWm4%gkmp$P$R_pR=x4WAp?G6fmgk(n7*Q4t`x
z^tCmqj+iU?(5g1^^YX&g`jeLrsSP8Zqs0(3h46zX9HkMkSz;OP$|k6ukoX&(>O7|@
zzCZx8#bZ6MpHOsA#fUxxfKv4<l~!9nqa3V+9YA=TPMt`K@*%_G&eUK6B&44&*ZSkn
zyCo{rZVxq!l>;_g5E<=PPoeg3Pe07OY%UbMdRuglZ@LZVpHNY?+F#|vjN#y(1F1xE
zh>UpZbTYV|G5mha6vmjAkD{kvK0JIvyO;-;tS1%UP?3;^#>S`-ruPFkI9}t(qo2sk
z5L5N!5;J33g3`+(<J(dra}k|p_G!K5Cw)#-?3rILU&g`UnbUj#JR<XOL9AoXiu--3
zM>9?(hoVK)<|<Mhm-0)TUH*$#lOhQ#`Yyz1c@%Gqu<7el36_L}<WJYYiO<uL=&L`6
zZM}B}P)H{^zrfJ#Oc+3`ex?JH+K5w(Q^Rj>wlf0L3Xj(#bLgSzY?f+ra@gOI4x@=*
z{GKiORm+#I*W3m5YRvXl?7#h^kTxwV^w&CG&MAyCd_4~o`fnyW<iQjM8XDKHbI!lf
zLfeqRWDi&5eQBr*6%IYw8dlu2*Gh~0?K9nOQ{EB6<Lgp<C~1V^;2@KOyQ7EO_#;>w
zQc@OjRwIS>gpfg5B5v-B`qW$z2r*y^UDhU+Dq)wHns?}iYi-kxa4mTd4lW8U!+-e{
zd*AhOK!#v=K)A*iWu<NS%CXU`X3cgj)7N^G+j#@qz)8_^*?%_L?{4{51Jjx5u${e^
zFZZvVY3j5L5@cT}1f?)~-r0}2Hzv+ES2Z84&0R7^y5GVU0br5E*Q(|1ZOD9bTz3kd
zrzByygtIFP&6rmLkb}XBk>OxxC%G*&7;*&7Yh3k?|D(XRBGN2ub(?x{%jagi4)1g_
zCIju+juR*wEzrO+XAzVI0)SP{ipF_w7+B@tMZrC8FZqNo&X=8)&i^E{#k%b3D*s77
zBY}%gO_NlH>~qCd_A`nMVS2rD@qeZu|2+~hYgPYIMZ=Lgu@&z41p|CH*56VU3!7p&
z(FhilG2gO5`4QJHX$8?_QDC*mSH-FbmVJXOVMaM%1!<@W)c9+v4W1u;fuw>(#%-S(
z++Ue?NoD@V>;@KNXyxi!EiTkR7aEsgOgyeU+kknn2)mNG?ZU{D?u@nM4fL!1&qJn@
z`1YdUQo?z*sLTC4&wlmsv%GH0AT^MTF;d+L|8QeyY3lqo_$$)NLkuz_6A0WJK+i)Z
z^}Us_u@OuuHR=sp49AjGOkv3c?>w<`7$5EW;hY%P#M9dXxdr@6NXYgq(8;_lDgSeo
zroQ(6))tS`E*V&q>gIf!->Xax1Cv{9ad3|Xoa_BN*4MstLv0<cd>hvW16DA_;Wur}
zP46{iPHXbGQSyO0y#|W~=j+JkM|YVK26V(mu<!w!QIrl8GT$o0$=DB9IH5>1%btlz
zz~WD-WX{ns=>9@b0&HtSjWDWOvN6(;UddX}WnWtZ7q|J5;fO2kob7p~+f&j#aCa9N
z{^x&FYudmN_nBCd4rdtbn{Zq_t<yzwyT1wxKlTI9D0zy5qz0kOR(%g`z-`zr3m>EY
zqE%oVc$5WWH}mnu+q#ZtF}YA(T?QWB8%)CHJiME)Jpo4g^&He2f5Rsdhx3(anEyxX
z37a$ANj)i>>wsBOqPp0_(fQFnvePH_f$^{UGBQ7wt!vp%q55fs-^1J5u*-es6oolq
zp<Y$FfoxbLNhdQy>=e8Pk|$kP!31>@=J^_!BYP~=6^Gk(D-=fF11RV5no8FL(LEsb
z4=rJGwn6*VLCCCa!z;|0k6djzUvX1i?Y=TbRTY(SvyV5u%cs_~G;n5-s&Q~i=%(<U
zxC6TnY<fjVVsy_s*<+~kyxh(O>rNx<z^ZH8@LM_`Rbp{#VT8$c(c$c`)tL6@owo^p
zW9%AzEd~nV<Jgpxo;|0ubmH(Q84A#{nfU>@%J~BuGcsl2Dv>K~ix?pjDI-V1xD%%v
zl<)MZe+N8xz_q-zt}s@vMuTj&6cU@?t6}r4j&gtHtcaaJM=>TW>gTIsTRM$8fxRLp
zB<b4dvdvmEVflA@&G&CKMK{Bs5i*uG&l>~q=`j0&Xnxg{u=Gc@xK6~`LN<2_^Td$X
zTl|>q!JxmPG>3(4(H~5Sz&6HTa65@>10cFO@I{u^lK_G2@P!K@i!A?($;I|95pu=O
z9m<U%DDlujFhDN}0_mieYg!0(gxSSXeIZ68F_@k_3$yqWayvfwy9$e{H!!LR%qpSb
zqrsa+z*zLu%~57owstk_m(Dkul+`7yiVq^jb2eounWqoE8u<@|JyGWXj~a?HM06)8
zPC#e5oKOIfs{%G1{o+{vl$vuJQt#9;F)zk&m0+4?lLzS&(B7_ZX_*4vr1#>Jgl@U=
zTmWuU3y%|bEGDxCof<fL|FU1V0f*N!1Ael|Uo7)~O%OK~)-~w^ql-a_5Ts;Qlwbpd
z;at9XodXnveO2IpE(@2u4lPmoa4^C3C|n3KGSM`LP?HV6g)NTasT}wq;gt`D&)0)G
znm6CTrE8o12o}#}ltg+pjAph_Y(HcRW^Ek3)2vn5ABb7tKLxKuE@V5Hpns`%?Q!kL
z$a73~?f)^j$+JU@{kUzZ^NA1~(c1#6{*Ru+BXJbHf3pDe$j(uZl5Qhq8;;B#tR1m0
zFqTASUw-RXbcKr%V@%4B7>Llz;_Mx9fq^<f=r*FKk$J|Ta>?%oj6N~&Hag2;%4`l%
zyTJAh7&$~d2A|yDV-5L796$8InEDQa6i`y>#>fM6q;!}-$?PDt0#OFW;@hCZAPTAY
z0P-0(s5c9}|HQVlSl2_s!ovyp8vSKtp6AP2Qto#207L`P#L8=%_h+?WzO9tZ>rRk5
zW=6MnO=kwuBZ=Yt(Or~9_R$lnm?4nf*}7a~ZUIitCEi?qucPU__~0M4;O5)N-s0Aq
zVn&CC{*=bK`NOyC`xRM<C;}B79RVY;(;3AH5^g6dntd=i2bY<gtGdYN!=Mwgq;#Gp
zQRrO!A~hN{Fenr7|9BUqi2#S3?98*32fynf_sip(2|;}9*`{XrKLyH67<rzr|M(vG
z4$Wk!hAFjvg99cSG0)XFZzd1V^2O@Y$?H>1$J2h&`C1F`3hy&#iP?_dn^(_H7zV2_
zehIyS-;*{|606s(F%Hi4MaM^nBnkfbV18)mN-&b9zQ-0ZI0m}13I7A7sba#xN^_Rr
zEZUdaKHp5#bOUTEczyuS=2DtoyRM)P@2Y=S{?P@gUow9dqS*rX$j8!pLA_!73mNU4
z5F<yWn13xjdoS}v5nLB`fZ_C{*?jL0aQ(q?!K1*#JW~*5T%~o<{*scKisU1MMJ+%c
zKa{<`(4eZJ!IT`TuTN5_pr@3wK#7Wmrd%=y1u4i;e-Vrn-q)9})MPm`jzv8V1k-CW
z^h&J)JKBiqIvi@M^|0u3=z6-t0+1{tg;WbPfQiZpRv_^o(C#j~2R@!-=^QK-m8TWZ
zg*jfX>EHbZB@N1MT%5r(p~ttu*2O7LV1S<fafqiRVL^+DnB3OLhsqZW1xD9SvHDH=
zKHIBxRXs0JMA%O=&-@)W%QvH^j-d+rt&a3TE;gMV^G6oUbX~Py3-C;ju@qZjVBz6;
zi=Ot9aFpQMy!Sx)Sux<;N)|SWhmE`YUs^lss5-W6>*Mb35IndX+(~c^5G;6b2_(1&
z4-Vnt?hsrPEVx60O9<}nZf|qD@9XaGb&v6lH{QRDGY(a!YS&(C&fi>h2<(KrvvH44
zZXF2pJ2_I0#))+{U)$J>(mGb-RDPHyudCCSBndbg!v~ATL<N2=WS((URb`ws%ExA|
zgm541t|Jr-RFGa4m~Vh&mFZ!uqc8<S^P9~TuGxoGWLK^a)9(a~Xb9wNrO?mCchBc8
zdlr=kMctqEcMc<4pP(q3zk2mvU?wl`?dp;dX1SA@*`z<b{oVPrADg~l>($D*NmWr^
zp4Z~{Tv$Y9kMqe#zcp6<8zu-FbKHDmJPO54sZMxU_+W;c2?9;^M-6M#QI(OA<xWIN
zh&pwZU*S@`jEu_F#oXfFUfyBLB>=2K$SD#@x28h}If>&V%Ce@mN+;w^8uS{Q`28`<
z`7F?kAQ`Qn_3~Al1ZN2oH961+OzBg*J12{?DJHD<=M&aHgdO9@dU7bA4YIQ<DK<Zh
z$`~@of9dOUKvGgp;;;wJGnJU*$=Wp#i~LDCq->Tu_%?<;tE$yvV`J}6f9@_%W8qK6
zZCe0w!Euw#{c>KVy_k~k9o$+_PZZ!xqP&-a?;dV7`Q?qW#Ga0pLFPGUvZbs3Aybfi
z+vdp-Zvl@sQ(|{MmuR%s{lxIBU#D^60MkAluVQ56%-To7_&#w(Vni}5lmjtx%6Qzq
z{{9e|O19*}tdCgfdt=?c_cw<g7ieZocD7xr$Q_rv70hiprcCW`7&aoVZx36WO}af5
z&D6$N-O$j^(z(or8l4Tb9_6pR3=9l_koolM<&N$o$$Cq?vU(*afOl0!58U+YESos#
zZcyxlnE`BWUc#0pC+Ui=TD_H6zE!du=dta&tornEgnt*Zagi(^7qTN=Q7JqeKJslY
z;B{aO+PB&TS6q}-JRqKYwHy`BR(04Kk^3ao*eLk@^hY?>Y?`RIo~-QB^{LjU=O%V(
z+!od^Uupx32N)OrP3Y7~2?@w55|k}1B8lw2xGDP!-fLs|v|?h1kN1Z>ea+xGgo<_i
zyV05N=@gL80tEWqs+j5$s^<QtyQ_;}(FLuYcYR9&cD&Kjo|p^t-dx_A6Q#;DGR@l=
zTcFU0?;z+(E)*~_d0pf!tNvhTs==P*;doU3VW8NIKT9*yZg4YK;GG8xE0+{4_HdD(
zm@!*tgT6&i+urUPe4`?E?{mE3p2~X<WhNWzdp4@2s@>=Wk3?<B%&F9<@%`Z@L{+=S
z9JO!#Gu>(Xn-F%S&tgl{FYfE@JbNRE8{hV_tBQJ@oju$G-&lALi0|S|-9LL<^fddI
z9rkE>sZ@2zl}ZEf$9%ITGi<Igj0Bj5$91e>5(;iI2he$p1jcGJw4qFI+D)i2g-k*V
zSyoL~?Yz8OPEJsgTF*3K{bYkAq1y*i;<0++Y^o}cC`WCs%s@>VprdQkf#rnyAlfAk
z<So8EuiVtUS7=lcOSc~cjIc?V@xS=za4}nqwqMy*>roH);D8`N9>H+jAOT_z`+~{8
zV3O76BjY4t%ukbH(9nqg>@PrbH1dlnV<s6?5uAFlxMi<_#=CF`KDk>boDI8ZIq_V<
z0Urfy)RoK<%nj5sW%Xsh*P+&{&pAlYkGgBFajMEl%eg_b(^<-m@d9C*_%N%DP-R?x
z28h=C=n?<2;r_F;{P&*9yyp!*LAXiHHCYA`U@ZDtQMq5qRZMG(zCs571FCH8n69>3
z<5fn~_yqRaV$eTof;UA;pm9Td>h!F22qdu977-K)h$8-06Ig{05k~0OQ!EA8;7^rH
zg#i0CM?Imft+tZ$ANn;fJeu4fF7m(B1lr7rEt0VN2dq3%R7fXSZ{zi<wM%7!FAuRv
z#Gg7-ms0t2qR1XI!_uk`0#9RrT>SDsSh3gX+;!z~BlF!{9)QUkY_to%T+f6j*ptQ1
zM?wQ_`mF%KA;QYsYLms6YND?!KkvK+X(pMNwOJ`5%Km5w)WxcMhj$XlV3?3qCh%|6
z@w$y`XgXj(FdxVFK^GLO^8xqS28>JZ8|a7ek=>v9iGm>rV(y~Pt7T&10`Qm$BgvhF
zcy3b-UH|C+Lf_R54mHL_VcaBf?xJf^epuAyMqb(HmG<<HIs=h-2Xq4wkDF@UTo44L
zF^^q8{w5#4Q+VFgdhIOqbOc`^N^GYnhaTe~d=`cAS+^pg|DaqP?xv3~HO-e?czm|^
z6mX}8ULLv@l$idYb9`r&q9^!kF<w?E)Gnxb=ledpe$##XK%LmbL!xF0*jpL@`r6*;
zh=iFiD!<jk)%QZIg~vT6u}hJg=9gOpP{bJ#X)rYHaDw2rNh(mHWn!)fMB~Q;`EQWo
zSddi+h-Lg^CvOeX7ttyGw25_r0zJkB^=}^0<@fzIbl->meoC)~grvL?*>MjOVip1g
z+O+_u&lu5vDj9O%)D)g7?i@c3pHyKb{gT|&i3LkYjwUs@(niI4l2f<guivD6i3fKs
zCVIjE9Z@KVo&m8TFV`Prl0?pGgI7?%^d&n4qR$*0tageqQ~rO~BVu@AwbowR;1lrn
zbTjDwzlF6g;a_WKYK(*CFZajz$^GHevmldr(0{6{Wmg1O?eMNlEdT%N>i=0wU%ML9
z<7}Ukq{&VYK>r^soNn*~^9DT8BI6xBmS8nuRlKAmllX<lLCdN_o8|mZew1;so*Ld{
zi6<dOh@IT3xR?rH?@yNePPgZn13T-tO2PeD+CDb(56ec;z<IGJH{?P|x&ODXaQ1=1
zs?5dVq{ye|B|MFQvg`e+_l!XGFzSFUDmVD0**g9P{vV*Hi9wXJDfEa4q^r>IX}tij
z0Oc%=Ti?eFzyX59%kaW+%TFI3l1OmL)}Y798qQ7}l7;Ctt(jpI96K{*KwK!hI=q!I
zDU;X4SlI_#YW_SbK+sc(3?CLo>%Z5ahP9J!kQ(|NfY?ClFU08>VX|b-!qm+H=o?Og
zKrp-qb#0MrFYvMGP7A$G_7So_eg&jIkO)13U0xOQgs2!aI_>j5-SPq{X*<vrUI(?Y
zX6q*)39eCGcUlnbFAtq5_VzC3GuL&>@GTF?FXa!e6yvp30q+XUeNT6zxO651za{<Z
zEnuWbWUkPdoMZSCnE0cgDNC(*Y@&%KHHXTf-=vDsv3+jbdA>e?s_t#DG;r`@2=e=<
z@-x2*g>=qV-+pU1`csKNNF&etp?twYpi|?XxnLyw^M~(lkbvUbxbnXieu<}pOXV)%
zs1oDa<sJc)r+!Pau6^dn=YnmZZSq#XEprT@5NB2xV0t(mf@oQI_UuW4lfzD7NsQ6=
z>zyVTE-ioW{2uC=x=n3X{1G0lN_Mjl@h!+CNux)3n6)b=>s2{E_D>867x)-0a=5|S
zIPGTO=C7YrmZ$Gt$PGPD*OjTL14>cBM#(zYIF1y4e1Y?P+a1`Zs90s<)~hu%C3epT
zHA{<ihzZ7$pRh=L<ij7IL!hW_Gu|oVuH;xmP~L#&jjE0w{q4=>{^dS*qspajNO=<t
z5JO5v*JjLg%|*LS{D%8hKX^Eyx&d;Qwm#>q)axh;#6qBS$lZ{IZ7iP?ATPMR<E`~W
zQ;}BoizP~(V0gNS3Xs{xQnjM`{YXkaFLA(E2ef1PKBP2{)s`=3bZ$zchlU~T!l*pM
z#NdhPP?lE?7~;1z5`MRcyEq|Q2E9q}3~8*oQu2mfqt8)p9C{TSXR2;PI4n@L_@NhC
zo8@R&LNIlH^F_eFc}9sTwHUyc=Erx{R@aY8mf04mfNNY2^&dC&T4E#Fj<(ubP6=Jq
zzijDWXn$_%+A~CinU*4F9;_-`ZG0ro12}p?FyO)S)Kag7k=6Z@M|G6|Ep~)7tQV*4
z_m{T2!MhPOU^lM$!H*`VXVDx9V~D22Yz!vh)$(z;{L|9^cfNEIvHk16n6@<8Ku{Yf
zo{jGEB={ytvHzt7sGPZv{43k`zjml!{V8n#6$1goiZsi4)<e7*$J+tU3Wvk=mvR<m
z+5glh{w5_QEF6@j7-ah~)d5n5cVWoDnv`rwUsX>UOVZ6tZq@3&p<x9v0GKF*Kjsvb
zmbNAqtR8>EWr#`<aLmy&NX|5zak~fIJZR06n_MZ&XF*rC*&cAa_$^jVe|2X^*!?7F
zciNs5?J_xs)APc#PFG!5_lIyguN_NBCo4;HYK_NQ?DEBhOGCr8YE%STCbz{-66cHk
z{d_+Ah2rb)8$Z%{UG|XJ65~=*-Y;IFMo@^r87L_X+UA828m5UkSw>`Lx?UdWUwauy
zNW35BdD(h@LF4K=t&mvm?&&Fdv$a;NPU$D<;lbOS*hq=g4Ds<wlg~F_C`e5m#8nw0
z;x<oeHyRomvg$#WkvjKS3Ek%8d<CR{jt(4uZNIi3O~fRVD{WVH9UTq#G%vyx%${2b
zun_F-(>W@qDF}m-Vzmh^^nizo%V{=@2d-hUqd~nO!u`0r(D@0t3+>Jc3@rI6_DH~4
zA>8bFNg^avtX7N@8t8dRp;NyXF1-W#gc_OXkysjwsXfqc5_Y;#@H?nEqYHcQmeEBE
za}|~<N|UVj#VS!ThZDBS4OLegce=W}2dNhfg%pxuY+fCg`93r}u12?dugAk@&67lG
zmQ2q63X@PNRQ3)o*!H=n7QJ4xb2GHA(`AY$12xYu-(yAWCF(i=3o8`}efc^0MT)G7
zLdFYSJ7^{x`E9Qt!0Xx`&nzp2%oCHc(e_5!-QS&3nsQxx9=5>osIte-v^9v_pD{mw
zzRe?s1&EE|RNufsgy!O~uz{HwAE2xw2)}03IFi2D7it8=9ew&yj?^iIw-dEf_ogE}
zO#8|DVw=p%m<^NVj(SWIqE;n`9)L%{Uwh)7>V6yiuI*@81{amqeAdc7h0_I-;q^gN
zLSi)LnR*CjW|i}92Eh5+paO3Q4cu@b$FmlOq`bC*X!C$q09<9{#8>{W_Jw`~9A=e=
zOQZAi@I%auXhc}Zf^|02&)mqwd`wMN&f5bDKdQApK5ROiWMzzOTyJK=43h{t{VIA@
z^E0CbVyP;IH7mVv_#?gb=E~^{uQ};9G@jjHf__z1oXMZC0kLQ#@C>+{(8y*mDZtV=
z-m7u^^5qX2nI9xPJFl?R`S&+7JgYaFV2y!=MK|#?`4k3;Oc>rBKK#Q7)i^1%6%fC$
z&VJ|NW{7(46{{02ed-JzhQiL|vqz9PfOg^2+uV8g1ZKJeI2SbyXJ*FM=52Js!Q;fx
zXw=2$PnB(n8OaDLAV~L1WCjvw;9-9Ahm6T02)7Hdi&79%+->C<A3kgzKZ=S8wN1w*
zd0%X1aOk%!(=m!O9PNM{pisl2eziH!j`=cu(U?{dF8%yNQfan}+QgNVhE4TWxh-&(
zUNO}L$#A%EaaC6sw8$wYtbTPUBN1|OqdDghf>Xdk;7$^YBqQ|s-WzRS(r~^#USl3W
zBcnVw^C>wlG%AYa`E%I!WL&0D!tlNXdv&CI?}P*dKCb9ByG72LJwCiOiuLo5yq+o(
zTJ&uVb~ije-ZV0_{M~m8FPFlx9C~+`vx93{6k@%U@4Q<t(LQ^9aEPc}`b8f41ku>9
zxfv#?!%1u-E$uGR(8sqkn0!)5QVT46I_-3aN#h*iBdd>wkH2gvx6VDXRgW&_C{;3q
zDNV=|GC4Hp=@+h+L+XY?+q<Qyt(eCV=3q8|=nHyYn3Tz1)oPXBB?}WZi|F6C%o{tp
zR%sfrq`NkFFZMH-M+~qs3V5C12;-x;WlgDKA-iWq_}Ys77HS@(>AW@DBmlvp$E56g
z3gU&w@Yku!g!DKm8;X8fD{E9OA6l-u8f@N%T#^bfo)n&YHX>X-<+$!_srPuywZ$l+
zI}2v=a!-0G=Qf8^zXR#RNDduPooHl4R;ri<!*w?OilmX!u=Q%<AK9(GW%!NCzj>3|
zOW~P;jT*z<^ZP9lCZ;ipR>cAow()@d_XfD%DluYe0V19l>^dhCN6WD7c_$|)SNE4?
zxbzVuoYWV<0N<h&na7J{Wn+UsMfpK#o%L|E9A|XI)zo^qU@@t$l*)@CiI~Nnw|sfK
zl^fnKW>tjfCUbvtp21tzHa^b7TQAlHJykwG1le5>N!s4;2x11D@BM1Dy?8q&8UOob
zNsvF+D%V2@UJhsf?z*eVq>f8M0!Iji#B#PXN#2D=9erdSr%VL7Y1mva88|YQG@=y{
z2%_SFGK+}VYd3!snE_}I`n1p3hmXS7P9KK*eoT*gTv238myNEky}4z;2GNyTz4Fnq
z6ZzN1@7>E1SAOa%4G%HLy$*IdzIA*oUa1Z{O|*~rw69)5TsCdLNQxO0M`$b4brCFi
zS^f|MFK9Bg2sB%eDvotBqMDHz_o&=k!e*1|?{`(Fqw?Y*7s*%K-m+Yj72))xd{bWv
zvx@Zl=gtW2C)73RFLT@r5U6W(>6ZlT8Jl~1M{9BHe)^iHT}#iMaLDHw_`X72{HnLZ
zlvVthLC^F4{rg{4Zz;!Z0g!`*wm&Z<OijHvsYU3IB6j(E4y&p3VAJ=}bGWw`7wv1$
zqCL&T4)oEP)0;{oj5K5~5+jPRmqs#){v{Zc-2gDeQ6IT@;WkYpgW<E0+-npNz@~R)
z*b@_I?(n4xk8~fN9&<dqqJqx}J~WGOQ4fP2vpE+)j2PGm91E!)9V3sw>I^%hfEe|N
zorjPWA>>L?^xZ8naZs~NS4J0)K}P*9aB5NMt>$wLF!9i<#(|Yqa|8xAlnI~dG<kdZ
zRX%4vIgen<m+>lHZ@x$aO}4@z4||5X*w-4I@l?G)bRv%6TX180{#-!N0|suDN&xCv
zjZTf?G-ltLKjEu+95iWJXg)QSp?*XMt@fbO<R0jC{~_t~i{%hzn+&Km2a6wRPz|vj
zua|%oARz&rD}xC?UR+e<)Ejlb)Yd`F<)=`9-h4K|JSeVXKCJ8XaJPPR^o}ZJQl$Cr
z1m<Y5^~!|DZ9KAdF}P?Fwt=uIkOqSU;tsI%Zj<9gLpHjN39K}tq&USId}ht=?qnbg
z)bGQ4Kkk$~zP`L}ee{MVCau(7-sk3_qs#ANsqT$R4~D(JY^7``f(^CZLGDB@)j7=%
z2slf&Fle#oDub=Io}`@AGGh$r71zc4n;@Bt122^?MMvKWf=f8NmMK6vkskYihZiq_
z6~`G58m(9i#b}E8Ay`|~Q}`Fu_-p}2#n|IeY^WF*rYRD(7@wEG?!~M2^JqE5u%W4m
zP5&m0PUiB+0lss;FIIi&-SixkunkhbA8dS#QT2;<f&e&9?cA7m<}AN6L|(Kjte-0R
z5mu77`314mXD-ekf}co+L5T!@xmhXIW#RIEN@1SD8oDqxBcp%KG#De&ya2BK+UvOc
zmg;v`=|-l{&HVZ9SBZ`!{iT~Ki=VMi%<9p%<^854*f&FW43(70yhpsD+Y5{u>MB>g
zlwt=H{ocnbUx`A<i{VY%G%FrpQLu)f72;|kNBc-udOa*p#--nOVsXP_SMTi@T!Y3J
z$U??D7@zQa-%~|NoR_@M;33|3F><*GpE1z#FpS8)ANQ=p#IVKKdq3VMwiiL)`zM0O
z3<`}qpmJ0(11P0Wg3*OYiy1cLxSL$a;r2;+Sr8Npdoa-vbz5%6Pfo({nSGobDgz?F
ztdfXl5Go2(GP)i#3q7Of;eplluj>{LVnLym^SLq>=u(lW5toC9fi|sLrsL3kwuUJ3
z5yvmtyne!#)V<ZY_$#7G6vNf+?MD)Sf$N*y@Xij&r0q;&WBtS5{Ztz<Y-g=r*QpNg
zV#-KFJdoUGt*joCZcuWrhqx>JsLBoQiIJ$`tR<|)u9kQJ(>9X@h5jS0VI~h3f#&O*
zuUUq^^R!q;%eFK$oycm4p8yXbNp||3+cy<D5U#!P^q+a=;`ZNp04!Q2AhlrE&m<y5
zPd>)f)D_^{Di1}m-!4S=L_EX5MhOC25yaJlY`Z9F9Ca{()f4}Z@ZXNy5+_+`(J|3J
zXQQDt7ApROt0(J*o$+9anPhCkM`wCpcu$u@vo{~ne$su>DjU#_G>g;2I9IoOa5m$*
za1A%c!;>lzGCO;16sTD!-!_Skfx+*v#D9vXCPL|Zx9|IKdX~cD>g^Sil<B)5u6Sy%
z$SsbD_9-yZ9l3cG5R2AZ_nMSd9zWb)h=@bua4mXYVsobYQN@CSm{*pfPp7UOhR4vS
z0gc6Q<M~TgL+^9nA?%Xq`1p8K22B53-^WWsq=@f#Cu1sN-iSD6UAEE9M;$^A^pRc?
zb;?A)T{t?kk0#V>z8I1VyM5^)E0p}=_Y+Ih?%T*`M5sH8>JO;%`sC?kdrO1G8jS6U
zbeiLVRy%06VPjuMCbMdy5lwr5wD?7%81ZFr)ThWhhxJ2}U6U%i<etJ}?lz+UC+qn|
ztgCnliHD)uX9_!ytU##V+gEyin<AGXZMk*NOUgkjLB^ou5hiEBExy4()pmGu{-&;4
zI}%Eh_GL)PWB%!qGEoD&okfaAoc&UPy}+P=P=zcLc5A(5fq^}g{Z`j(SBN$<E8L1@
z<zyM+T%LiFZPCLb3{0X@3ijB@tsO7cH@LXU0>BZO)@-oAC=~Ieen}w^GtMJFwqH2J
zkwi8u{bj;l7ewak^UBKEi(DcF-8BD;`X=bL?{2>$vQ-oTnnzZ_VS9`Rh<OOjes6ZO
z3wiLgWS>lMBCbE(VW2N()x8FQ&J$zwjhaE%#>)RdBEaUn_CMp>_3=jJe+hX1SNQh7
z0lWE??hOOFH{Q<(hpniRi3;QFa%@=z7`){esWq^fc+*4=WJqU;hCA1#6WdP)g<1(#
zATXwB4AJ&9G$lwK+z6c&CdD`#su{H^c@j}j2G(0Ys_vB*_f7UWmL%pD2kQ{l5EGpX
z>e2<E4|vHjWKX=vuXMrofRnHWX)G$Bk7XK2xVY5U8rVZf#Am;ex#+l?b*zKov?&n6
z5qYUsTxGnhF9-<v+?LvkMn;Sn8R+OT`QN_-CSe!KsOd<b<Jm9ySoIkB(6CS{R)jyN
zCNw<W$1;ZP5R+)@ff<*jBN<S<P*5o1Pwx@Xw{DJWs?ji^sj%G;*#jlw);U7mhF{2L
zN7F5$>o)CDjRakVVawL$E*$|Xty)#a6d@k3-EyM!H{mz>uVXb|t!->HgA4^&=ohD_
z1mRMZ62>6Zl$|XR8aiF_1y7Tyz;Ei~)(1E>*d9*oJi4W0x&+-sFACSYK4`0Ccdmvy
zi7GH00V|8=%|)C!r|7%3s&TC!2952)>B54%EM2XdsTDp8nrfQHa`vE<6?1>kZ7DJQ
z6y%kzYysI(kT>JA|Gh^+V9voB*uQa_B9IRCK?titL(4Q^_in+HJt)MrWUjz|@ssP~
z{ZD`RiU3ZVvPhj+nw2?8Mcg~5C)B9c)8eiopzwlH3^NE!*#%=tBE?+s2NHw^TRH;(
z6+OkAh>+_iRrQ4l?<5SxPnD555&5NLak|L2jXWK33bGGEszf3lmMyydBe^n+keQlI
zq$-k@g3mOS<2uh0@w!cFX1ki1AVlZjB?=tFu?qGoiE_Db^Xqe%f;=vlx%*wKv(y-u
zekOBu40lVEjAyobtseC&s9pnN%f8Ue`S2I<P=D#~9G=um?%jj1z`M=Nv^`69==5J_
z)1%G)YDHu-d@Kbj{L#J7D~fKMR}Fe1;8!{G!ui6b@+~$i7JZ01K1~*m7FtYak7A42
z&**F0euz|}Xw-d^|IIV;gSia<uXxhuO_=FHB%a0lX~i^Td$hW&=qRZ%Zo9*qlmRz7
zxFPe3{WHzGZu71v#YqZ;xOnpUEdibJ%vx^32758j@)}g}WODuZ%s_yz;5cE#=x0Xf
z6lqe7<7QZTD1J{qaLxG%y1d{A`g>&=$Fi-e4WBQaQe+ZnJ*&`dIyX|5I0n;8_i1@m
zAG;!6;k-Gf6vD_9_H{N|JUY-<!l;(*+nsC~e0aP0bwBiQZCUvJnSQc^tMkwgX5FTB
zC?(PjMPfNcU32blzCYEGr+pSZplY4H^dK4-lrut*8N#l2y$^BU)AeL9_uqJv017r2
z;c~W$FJp8-7cKH&lC{sZ_d)$ODt(t6%$j;PkPrBAXy}{&9yE~HY5rJkg*ibf=zhEu
z8tyE_nO_zu&CYHwBy>BrLM_7}<Eo-!B7Y4TZ-f(DfLQcd0_)hmnGRBv3b9A<u)`<&
z1A_gh>5lvz{sa4V-qaR&Ax0h$iz^aIyG`x^sqbciH2??8%Gf5fEMFZScMo_1KlG%}
z3AjtUv$=2D>Qb#n9>(ha2-i|#{w*XZdvk$eEC@~NaJ%5fcvtVfIQUc9)uoo1ExzUP
z+P=#+-+Tk_=JwkQ0m#f_NBr6oJ<i6Q7%l~4?kZ*<y#!Uk+7z|e9H(<Acd<o~<VHhB
z$C{WF#MAtyEvU;B%R^RM0=<vVR@JEVSI){wFiiE5ugCQjlAUc07ni?(z&Aa=S@Jr`
zKJA3V1(*-BU}Ap$k?f*l%kp3~HmbXDxn?-0YTk=+(*`Hk_kpFg?GCOi{Uvw@Y#2ZW
zfAWpZm{7w98=Wpp1}5XPl&FFf8x&pmG_`f|1IvG>k+(7X+MXaeoDxgDt;9z~4I4=M
zptOqs#_ysYoa+&WmKAATQUOfRh^&3}Z{q<Fo+*Rr0R1_@x9Eb2K!BC1QhFvF)@ia<
z-XQ)tWf+&#@_c`;SXmh(vNjMQkIv}SHYJvL^S!_kHs=PbS0AJdv0*J&kD5FK%s4Pj
zyB0P$SFgNk<6vRt%g-j?vm^E#Hef1_SxjeutCB2$ij^O3IX1U(bZxNfSz(TU{c(AF
zefi`xxMe910Cs463{9`C>jGM8Z^V~qxdf6ydYY&W3<uUdUfoC;9kOnt5_G-p39Zqx
z(d=mOzUY)jB1a=0aa_ylTN6r-u;#;~fTU<@H16W4lVLFC_Y2s6XUCYuAvsehWSOhC
zbA5b>eL9P&=coFLJGNPJxX?O*(S%+|NWw)Ex}lLsOgtoMF@a7_DP&EGKYze_6Z4sT
z1VL^~8ls4=;~#1Tbq$KshzMajSw?<xxBU$0e}sC~(&+I`08HL^dmFU5P=obs7v_c$
z2c;m29L@ugo;L9<&N~iX5GKGjb}j1wAyJ=}7e=llzc(|IAk@76#UXkZ?4*|{aLrb$
z2$P<lt#sT+y*u>M_wBm;EzG0SAp>~a`Vd#@wEAXwWMrp~Kzb#%LQ(#<)VInHQ$4ao
zeh_cPql`#bt9re$+LYl$amZ0uXicF?rhU<D)mw&azj!$cgnDrlL^`Fs83<zd`sV6^
zkp(Iz0Ux0ZK0jU~{UDpauqki|0&TCR<{5gTBcgyApj)%qCs4_j{g4>xnUN;20qoMY
z1`s51-dJ%i@K!cBzSD{Ri2OW`nfE}rN5el<&`rDo<T5Q&7LGQt1KPoKdh^30Hg-`F
zd>3|>*x4{8<AHQUWg&qX5HBR-ygobugYBpI(I<n+Wz1D57MLM1$nS7!zHtVTd#Z@?
z*bH^_GKdp9d95ms8=9^BG)y-Vwd3!)9^`xWh!IbSYcF+ygg|=nRq58cMSX->eW=|M
z$`_26Z<AAF$#r+ZwxW;@i0Z~Y%1@uDE(P`KGwK`=buO8d&Q=9d_t%Y2aD(#QYYHvM
z;nqE^qt1eY%wMs_-XAS@>S!_jG5HJ|ZoBUven%5dL==ej>uH6wd<4Ohi%q|Ax&G}X
znG&8l(g0t58-LoFR36w|G@~1dtxDo4bbzFL+uT1Q?kTC>>9S@&Ha=?Z%+9kjjs%#Y
zc#jDPB**NGT060X#eZOU9GirU>A_JmNEufv>m$57DSGt+5q*b}#CqT(9o;;?%ih^S
zONi1XW6q4`0G*5_gVSD4kC+6tEb22b7Qi>fL%)oc-mI9R1O7xyi&@S)u0Q^Zhp>#g
zQ+3AG+PxncyloxRu66!3+V1;G^)WyKC$wk`nL7%)UneBK?jS8L(vTFXf~l*1xkpl3
z-Um<WxcL1o5VwuX+31~30pgh0T|HF`*Yw$9H&_=`K?<j33JNj5!tj}#^@k<-^+-+f
ztX>BhfVUHACO;PT$>zz?rDn%(kbav@dbiY*C^-RTzZCoga!=*cX+SPTNiaof%TeH>
zXk1oi8$tTh#4&~S543_gSW>E}?CEU{X%dS&D%n9WKJ#{(V?G<l-p7w>A5pIa@qas;
z;6IPhVjLgSV!=}aZWMxXm7<#5s31@33h-yK@}L>RP}Ss$@yzf-WS$@dge-oT_?$n+
ztkyh4hL>mm{0s(Fw1;l)xJF7&LRgq^uHVK<KHx!v-YbGHaBRk*#SR?Jntr}k0FQpi
zCi?1iZ11W9oCzYubX1Om6bh(c1Y_p>+~}^BjI_9T9x<Vy9v_n@p-;f8sSuuONm!_a
XBbZO}Ge0%KU!dfqlwX!h8u|Yhg1hSR

literal 13675
zcmbuGbyOT(yXBMM1a~L6OK^e(cZVRsgS$I4(g7N`#@#KrLvVM3ySux?blz{~es}K3
zTC?V#Ue%|2RjsNzwV&VF`zer;f+R8`0ph!N?~tXX#8lqBd*2M%{saF3@>wGutNiXA
z+B<16VKw)(;}ryNOqW%mbA8j;!5M0s#p^t`Y#2nQ_&+MNI|VRcRV-4S3UypN4WDnB
zK00o5D>oH_&E@4sr-vm)v?0rm@v_WKXEjRuKBN{YPf3H%?G}|r*)D(LNzQ#zHPT+c
zup`*)#hsnqr5g8;Ic!NYNE9T{edxu&7Y8ATz^^RoFekVyTH~a5o1WL|NJ|L0<b~wu
zmW>{=VyudziR|ORB}++Xg~YS^nu>CAC~eDpQz`2naBv-w0iro>>7s!YTKsc*Wn7>+
zl)}|u^ea&?h2hDZk!h*2N8-P#4aF1STG7Ch0l%zo9y{g-1vC+l?NI*n-I14!zW3C_
zo~S|$K0ETF^<z#fIphcEjz%$fGGSI`bP*}Ea9JQ42K3rlh*Bm6hO!Yzg)v687U(|S
z%d1rgc1d|qFmte5&b_|dTSJHZ^$|zveX&lrgKMgKe~ad(KnOZK`aFpEc0w3uxl!`q
zY`h(F9$&J^&OfyH=C@m4dHy?xc8{o^qr9`srtt!Av@b6YZXdQVvfE%}NcJf)GONne
z858120P(M{&wT#r4v+)jfsPz|ojK>FX#zdDPn92^uFiVo@e(b`N#!JaAAb(#SJ!hh
zUbKcZ_Ve8{Fo0-FSH4-7Fl-MY<q_)TqKR<-PGw?-@(|CUw=%j-3h5-P)$F-KOXA;v
zmvlJST@=S8$WfuhK#$z98k@6Q0N8&Yz8fW%lLJnSH!f_l$)`p0aWL50=MDGPFcGPN
z(UQbQ;7pn&ol;S?;g6WvJz}KDar;ps)0WwqPo=hOZ=r=(<Je-SYxCmCprD|h-4kW!
zta;qp?x&Mwxd?sx)ZZkO;}`W8mW?Q1@WtzJ<oFevrG+Ag)3V)waDCQ-_0tBEb)3`#
zKTM1~wwiEedoM6>(b45gM8K4&Bt-AlcErjy0_qO(V93?B^Lnt58Bk-zPq++qJek<Y
zDTVBoJ9?z@)flrk&SD#a;CjS@DQWuc-cFU&)rk+Tx;Wp*SiXgHeC=|IFbx|sXD%t;
zlM+J1$6_&iW}ZnS=kCb5FVdRoKfLY>NDrpBB9Ic=E{qp&zZ=Q=ip@ATeI%&cHfZXG
zmg#|>2x#5=C_#S1*rLN6sNV3h%^e(8|8Y%JWqY6}#IR82R-FSUX0|nz_i*(ixL;`D
z6?gOZh_Gsa`UV^#w?1(qM(F3cFLGs~CHc9gaF58*f8u*|j%#5WqdrLD6LV0qep(n;
zv#8LVvbL)X>p7K&zf_TPzTb*s3WuK&gDxfS5+NWU$Tps{tlI(4D1prbI7>Bbm>4Rk
z1^4qh(8CVUMd2RlcGgb#6q+do$R1D8i!BVvI%BdCn??*9RAriGEZ4jbNBEBFIrYwj
zO;mymdth_(SJ)?^iwIJ#Ph{}G?OhuNEjwtx%P&ODI$t6}Wbbk1iG>cQn`GZxz8zq)
zwQ1LM8`Z^}cB2XUZ64{KIys8aYShSXQ<DE#Ic>EeH{QwK+6(+JgHVOePHj@YSQ~sl
za&&lUJSf;Qg^>*+L`!33dD2jr_02J^=&+Hb-Y)1fSt7zw)CmnSd(7tpXcF~nMKGG}
z@+LEwXKC~Ctnaror@PmX-~BSaKP`V_mh&Q`f0^uoZF&u;TAdx3ltH%^iRLy|8SHAg
zRE^CcuT6PZN1&}48Q2&-XdH;K1vOm))z2cA{N;9-Wls2A-O_Dw@NVatgN@G>T;{Kf
z+K!7R``hiF`GK%rY*`JPD~?qzDn1G(P;DD)#ai{*d~!~yo8wulMU}GQO~hk)Prek#
zkNvE~E9&5WvBhI9EuFmK)949~-V4N*AIhuoA?sOhuji!_aZNTK-Gka&_v+8lkQtYV
z#}pH6fB`TG8=ppqczdoWHN0DnmU~q+BO{TT+*Y08ziHZ9a(Q3xz=otg;cWLx27BGY
zetuke-FiaFB-6W$E7nhY8-M$LlehY3ezqvSL@({F^Qbg~RV!86%|Hv~Q0AwI^4$Ef
zef7uFq{i)|(Utkkbh?_gd}sbbH`K`A*amv0G38Y<(r%@8)<MngN2e65de=gVOo%ow
zD-UBwjm!4Npu&{zJg-^<ZPz^@v_!XXUo@QCG+iOfbuq9`E;sk9*-@!e4FDULEomps
zYaGzBb6@>L^P7W~4Vf#A!+LeTll`^v*dHYmld8$<SrxQVjFD12qfxf_CakXd*+L-d
ztvsy|7FDW{?qV?ZtfGbKS&x^dU)8$@8aZ4^@41B9G^aklZvZ%P{xM1y>z%&H4}%hN
zqkzZDMC3Hp=?DN~L&)dk?!G8#1cd85V=9Z$u=-4+5H5GAoCWL8`6#md_WQ>(geZhW
zs9FdNszPaJ!%{u%zRvHT<>N~Ac<O}TM}zaR>);~O1N97F^U7Qu@G*!cCZ;c63rj(F
zLBUq~WB8aPcpcg{D*53RR=afQy)B9f_F6Wr0~~ZzJAmEYxAE#au6}UT=okqnR<+~W
z#+1D0*2tii2PV6aRvSVFrLl>H_=<n7+1?u0Yox_OMq^rfBGf{Q^=RZobO&gG!m&Tk
zcF|$tN^z-iE2~}QGqun8^>R3r<ap<}zx&qhBq~?3WyP3`K|LV>a$QXg$>V0r8^wz8
zPEh*YSSB<26ccRj)L>_Y!Ki6A)#P}9y<AKsT70GVtaL>cD|c2SC1E}3r(XcwJC=dK
zpI720d3??<94;D1ykCg+1aGAC;p1hDRsWpkiF4#d^$Bgy)vSVMs8h>4_uCZ-1X&na
zwe2hd_JQE+s6W=0$HwERiU2Aq>fv6e8z%J_;i~o_FV3&%NGDe{xP`qDKjEr*M-D(u
z6PWF;-;<!+=F?l3v1733-4`og^5gQqAF^jk{XzL=5vQIUMFYl~7Nzk1_|0#|z+@iy
z$<8=J#<Yh#KNDaj)BFSDb#7ML4<`hpE=^Gp)aJ>6=N4R6Tvp7T4B$ugH>9-x`ubyV
zv9t#2&((c$j)|39hrWk{R>XZ_=1}uJZcX-68M4aT?#0jZo!l^6z-A|}q9D&5Sbq&J
zpW%8f+rdC*J4|u;DtM37{a{esLKL%N)!T9U9k!(P%uW2v7-#+Y&8YRSF%{(>b`!kX
zNaZ+jzjNq?IIOL$)t0Cy-UpX4&bw?641UBZ)h)KqoH;%#^aBi+M6`_Q(=+@k0}q$S
zOF81_awlaujK6nNTu>yts_hm!U$1gnx+UTT6n|(_8~9D~A@rm4D!_KC?&OKZI&zXa
z?i5O#swYB?9zp3zn$k#ya!Bf={-v9ANJ!qXDI|DFD1wc=`Mu#oVDH^%DE!ADje_*2
zW;4gGqeBct(b8_BTw-XbnDd>%YSPkHS<XVC#moMr=<VKzia#7LkK9q*sYjWuC-Nb!
zz=?QHxK}2#c%|r5fp){EL2Tj^p#dRAYRW+)?iY`TZldwAK~G9{i^KgqI;G5l-klD_
z%oF!3o!)F<C6?5T6}MSBv$kw3039c-3`1J#@hta)!D+DJ?OQRLUQgdMm$ZNyV#Hx*
zbr!VxC6wWwP^l^5dH443Ek`_p!=qZ^P7bVREGnEZT_O$L?|b%N41M|8$cZBT-{tf3
z`?YzN7H_eCf@{}nxJtYmv9<#4dq1t|H!z%M6T&o9;nM|f=H^n=d7-4YJ+`E_^EH~U
zVudg%!elEAK&cHm;mVHbF^-PP=yKQ{K4q5sV?=G~kTSn;aWmM+c@BYU5*N9mT&t*a
zJ9`}+3#W|g_FcZn8z?!|DCDciT2yKO!ugD`YNVSKUHM{EWrjJv>vZWv-W_PdW;$wm
zN2}?e#kp)CjJZyVi@kA>5s18mGLruJ6$H0SkFAor(&Y3et;qMBwY$dkz64Sr)A_vy
zjc+}_Dv>7;02<eCR-M)s0q%-n-4@_blf3UUv)tI&8bQIZol%o(ih>r)Avv)!?)aYs
zRng}3Ka9+zJyD_IWZ2%;EGOOJgj(EHUgrN9l3ydHT<JM#9Fb3+segWXi0EfIvwgZ0
z+4<a9VSB>Em*zSkqgiBE3~V%}6al112ijLeKMHW~*;S%tInRNs$O2!s{n0urzYoZm
zH|co6l4F0fYH0dM-2;+kPMoUMBsY9kr~j$)J_0sOSPZm|O#*Ae1FsU~k8QxY`ce`#
zlka9}c|>7(w3+jR7Crn3M0T8|Q9gB$;vygI<^SF?1h!>#)^TPMMwL{H0#3BOuGieQ
zp&=4hG^^Zy0b3Ift>Aa_@4TEogxslbUK{z2iXZd-T)Oaqyp~xxsM&yxs`EHF7)jh@
zTSn45F{8MEY7d9vY4ccVO4vX$TLbcb$9M^>)M)dWgKP=r7H=(P@Mnx5?)oU${e?~d
z>pBU3ITL-c<&)dXA#Tkl)i#~qGgE9Fbh83zL!E2jj~PNpCS79Dd*4++(695Fv*36<
z%&ae`Hjiptx(e`wj`hng<Qw=ecT(`Avj#<EE65Ter_Db;JlABO>bJW9yybeTrf@;y
zrbEQ%vNz$nZMX(0bZ+I1?|&q|^M%s?wXk~IeTJjA_gK9Gcrsqn=hrO^o1v+AoptFq
zYGta+7lLL$^6UDgpkK^Bo5-zWY<@RB?J@%wY(E#qt2dJB5=lX-W7#^d;*qzp)Q^gD
zQWg+F{YvMuNg8xTq+eT#LBN2@-GNw1P==I>wvCK8X=!fFmJ-gDG*q?3gcGJxEG0{v
z@x^-n+>ov|Y4+%5hVzC~UPys_&ln{bi8u-E=KiI|P=aum!E7($$oy?0QBfg02n9<H
z?u5?~$CN9VT;z*~pbaU_7CylTQkJE6sofuUw;c|~1-<uXra&;GK{-RA6f}{;*rD(?
znreAoC53H`)AMB&^i28LDmdsqs#C?ykM;5f+6WmPmo95BH*5Ob``26Ss)h#sjpN^9
zgk0CG;U@sTZL{n@NU|;o*=TWKKW`+rKXf`K;Ra}#jw`J1Yc_ix#bU!``u;}P9jg8l
zjK-2q{~+iz!2u@UKkWNzBIVfy>=s`dK(2J`&e(^y*~qU>o9u=KT1mTUT7U&sHqF-Q
z8aunV69`@O5Y_I^xw**rBAiJpN{2U_omgj_t}os2p=icOJ-6%Ze8pQS&#te8vaEq#
z>BX;WFsE%?Bd~oxQEb^}59R7k6RjjjvO&LQ!LX&|&pU@!+<Wf0EE^-9Rx@O}^X4vn
zVHbl_aB0tFJgy>m8`6X2&15yX;%pjHCC!IfDBmcIGMo&|miZEg2t)%+W@cDRXk#3S
zkT)h%NKZYu66&;nan{u#M<DNo`pTunUM|)v^AE_-wi(Pe=Dzfc2Y1ubAmL{G)=WSn
zl-7PaPUk1Z<4()qD-kveiHbD0h+s-8HlIEa`mW4sx9gtd6x=oO+TA(|gFA_D(@pC@
zkUcakv*Pr~ylTM4iJ!u0S=VWPsY>fiy~Vr|WmdRjj1nkNvD!LkUJG6F@|{7F#9{`C
z|9^M^Bv?75rwwPkI3yU*4PC=K8BVsaz+iJjeR?50Z*$uw6_Kt|aUC@tB>2ctJXn3M
z1Ee0=V3rfV!9ROuZ#FZpfDAe(H&#_6wmlP;ppvVOpC25dI#Bn10e<0edZUHA1Kp;!
z*@&z?=fL!%zwNt&n6W23;-x7>0;TT`HZ2Ybqr5XRSe1Vv$%KvbDl=XK3;sNc;I)j%
zq?CU%PdN2ln!>~Vvc0eXu=&2Xma6wzJB8DijCEg2+YN<!nT(v9=z)m5y7AQ7=)SVw
zwDsGJGp`7y+e)yXNFZ(#>f*;#2H5Z^IrlFo6Oy=YZN84Cz!sjBn@5ByeFBbI>_ZkT
zwaT)%gb98<_$k>$nZs^TEc&+^EcsybIt);cYMXFqrGlBL>QsJCPV8mDRif`PnMmhO
zG7l^ZV;iB7$#^HqI6e>`JE?SV3=fujo@N>}$NO|y!rVGzO8VIvjfil9;ms0)2?74~
z2tQOL^N@g}cSQUv31AXra?b`$au$}+A`@$(?~^)(P6V6X6uRN11tmk<>*KKI<9NPE
z`*kpDLXhRB2YVmpod?%(b=_KLSI7N()-C9daZVDR-=3Bv*Y>ab>Qe8<(!T71;Pzr?
z<KH}A#_C^|0U8-*sD&S8BQx5*w!%q{mW}AI^4@}4`h`wbvHMT?+8wqg#|M$2ggQiN
z5I11+ISw=q2~+g}Zw#rw`PK|}U35GZ@cb7tM8%hCVhdI?F23$L92^L?tbyQzg(q**
zvKxf4P{N2sU3D$uLz0zl{OWJ0IIbG(^!*m>@h*U=PYo^(OBppsAzrb25xoxplsU0(
z9i@nJqUIA;n6+|>lK|!~<4y2(4SXrg__QKc!KHnz{H(MK0z<sJ060rkb>yEC_c&$i
zjzTc=@3#K%ozu8CS4L*@(hO76P{F`i7(in#rkB+e7dO?`=H=yWo+n#c!P))vz6vy%
z<9gjVp|Tl1^*t)FQK~Ibmx0f+(s%^`-YyHr`ya-F8H`mdma3|(-Vho)(N~qV5TQa~
z6i(k+8R2EP)FS;DxHmi?5K?c_ejV=`>VC{bsuYoIad{Ksm4BIT8n$n79q%8*>O`jA
zKRKP6uHes9Otd3kF+S;xT8D*0KO%{-BNxcO`p|B7xQcn~Z!UHF!&DUeK$ta^{_%iz
z+XY02fHc+Nt=w{mdZ{KS`>CR$Wk!ddMj_$TrnnI=l{$NEl1v%~x}XduXP3aGibHGm
zZpi7MTg$P7>INRo(iSYHDqX#u!yxuoJI+pm_PoZ9B6~2G*hf_XB?E#8i6o=)GD}|F
zY0bHP7qc3x3~zhp*8P{4t@5<M1?uSeKQBp2ysAc9sHwUN8g?jV4C*l}E>f=5rG$6?
zcv02bEUjkCEBZK<1V@ft?mKlS3S;jZBI`>wQh&J9h)W;i6l&Q8=GicQ9!QXN%y3j}
zJoC5<A6<8@Vn!G5K|=PCrWHTZ5xB$fVYir0Mw0k$6G%nB%m{8Crw!09DH&7JF1DL{
z_?^A_<c@uz#U5kB*-RT4QFxANpd4&lFBJLF+oFh_QUT+=I3trw6s_pqa#_uZ$GFdF
zhL%|&J|g1ytcwSpWvR1T+TQRWGQaw*>8^cs;&AG^#G}yu^*MQK<GaI=QJlolp4wD_
z<$W3WLJlKxvR~b~=M~oGj(7XncBdK1PAs9Cn`K)mW6hzd;w$!-Od>12y}aV_^LJ5*
z!Rd`WX5&IK0KjWH!~nziqbwQjrAss55-d^ivS6eOdc>Qx`03$QxH31Mf)%#IXm#@N
z-+jKPuSIr->82?OGI#)G4LFsRxi!;gi>mk&1BvS(!GHGH;co+R^pZZ-eRnaEpo_Bt
z>bo3SGT+hZ&5p5T`fQQpw+p=~80e5s?{&OKT9I4xO8I^B&wYXB+kz*fgzuvgfky~%
z0HKKlK>h`c{)CtKUxwE`5(Q)^bIi5rP2+3tgLE>Xvr}RF7qx<uw~l@GpjtmOd!ZOp
zQK(lXP&+w+h8Nxht+BH+#H6>wRfcH;h+IWUrRBnhGoW0kjPvF770^V4;A9YaH1qfs
zWnX>*5c`mS{zw#ndt66xf)a*D*sbLV<@=3iq}X*>z1QJRb=!%4iAcoeW0pl)`j3yg
zg#ocwJICBf**vd9o}WR9%wR_SwhR-W7d0QC(%mH;Gh83)1T4%kpSQ&N;{rY%vixJt
zrc6Nj#iA9<&Ti!7(f0!R;dFMhzUK#2OrVnfrm7h~ouKFRIpxMt_fRE*@SK!f_|Gz(
zX6Ku5{JXaEu2OXL?5RS<Pw?4oJz<BAG767L*JN1YGn^bOTFR`Z(-YSlP#eREnJ>L9
zq`$C0Vqv?A$cRc!I^{2t<916<&|Ds;lP&aChj6{9w9KX8VM<~DlCO38DqVgJuQ!~+
z8iPYJ!B40{&mT_jPWX&!7A`NTv9OLI{_>hqd`N+Xrl0vHPo{z!E<s|nGjDRYhTY;+
zk<~**lPa}w;R&xhy&NGPKF;$@3b90u$wnESQOc|L&MYt~ByGe{QW2Z0AqmS7w%W0~
z9twNrhB*Upe)?Rp*mvR;g?_DiaWRP!jA@&}B2OJAfhmP1A{Z8l1ZF!8dq`rcA0o(9
z@WvrXbZi>&VTqgN+Sh|<xngdoIr3~}hFPWa&9}Q#^JeKeHrwmU0am6Ss@mgaz~y{0
z`S_zN+{e_f4l)ypFt8;X;EEI30r_O4HMK}3+|L-&BYqQqLJ4Swh5K=}tz8SFw4{Ok
zCMGT~lT)~bJCQ@FJk8DZ7sH%O5!Ayo40)8jE1!KB^LW%+dtP4m8F0^SKO_F8-T75k
zRh3)-c-EE5moqWp6oz+^wvQ$_T_9~<rLLJ#q%*^l{tj1e)tYpurE&gJdgk3`eS#vN
z<+(bdqqv1yEVa8G)uN+)v*@M&SK2drTA5=)X!Z$AOu!f*%~E9YUlkmdbpUa3WXPTo
zDA1RP6*Alk<!yXaeR5kQj}Bs=M9g(WF7CTFTLM2kv*?wxri_-)&Iggng4@}YP879;
zZ2L_cYKZulxS@6d<t><DGbF_L)Nw}KHNU|O+#OaUEoZNEbaWdlZu&oC4zL+!AyYB2
zF}NlVqogjSW|8P}Hk+wsMc7-^DwhqYQ=~CxBqtr2$!}|Xn#8f%3#c(n?6sU~z++c|
ziy}^Uyga7DBN!k7$EJem6WQ5<=vviKKq&fpjXINDng{k@Bnv1brAZN|#I|R<rPPUi
zh|q_tS%0(pc%B|Fcqsi~)@`;NPSAxW6ANMdw73)oS|U=Q-I}(f+Rks!WYffa*PEtr
zHibR7agU8W=sRLvgwg1)Dc5Zrc~VG*9Km>$QQ()OLT&O%f^Q8$UOBC_!h6+BQKf|Z
zCf|!s>%FgERh4pYuL(5Sb)et8Ll=7qvazzed9)ak>W9`z>tDa-JRWy-Vf!oG&JO)G
zuPJ}DNQ{4QrU7s<bpA>Pi*%r()nH`XzE{_k_TW~0w=QY@>EAMCGQOe7M-_t!j5e?z
z=<pYPgoQ{=_ZPWv6<LU)Nn>@E<yN%LzzCgH@7bpX##rU?pVe|~s?(w!5>11n!5HGv
z-5XN~UsvF_s`t)YKl)zOs^wp!b{~aPO!JPr^le0U9)+*vtmBSX<20fJl>mW1{;NFx
zXVn}FzoU%_$+0L-Qr-lK!Mlt(k)!u7T!`g-NB<9nM93(JRNA8jr72tDLym5{PXPis
zEI4JA1Iv#0F$2&=Gy-Ekpy1GrPpJSBPB3?1B6V+x_5=Aw+VsLuvcHE;{eg!=IUuZT
zKIUcB1NqnkTR#NLt>N$YAO0hZ|Kv`U?5n}yBm(3@fX#eas^P>6Gp&SAz{gm<$l;|z
z(=!3*=BE4#?+jfQ&y%`DN2mV%8>Xp(ov~#0AFt2B2qHx+2Rmc;MbmkZl%i>ein>qa
z8)Sxf_Jw=tyKl^dgiIVv(V+cRPH`7}Y;-m|)AB`q!y3L>>s7A}M;H#%2qHEzKKUUj
z0+*ot()vT>g0zv1L69;j^UAw5;xs&`rRg$mA<c}Klr*qlk7b>Bkd@(R9gG))6@l;5
z^zX_s0jGjA7JA)_6K2Pc!~b`2{Nrh88F_W!?3O)~FE_N@ZqPhw{9v>*8+!QIVkU1y
zx!G8B{w$r`uc)unY?-PlC-AJ#E?`B!%{KlGLjp;^D*$4a%@(Cz_f^auM@v&p(HMW-
zqS3blKG}Fh*{=A+FE4Ix<r1UNEc5H9$l*j#M}Go(;}GK$kP@RboJW7!B+Q;Mu~x$q
zSI%T!>uJ*I-QmP9mPVpbAWoUcn!-K)@`1tobu^(6z^lJgv2Tvb*#B#_77pi@<Tf*b
z4-u!tDs{cBEgRob#nI{`l3F)|{!(IeDjR-VF;O<N5oK)B`|NydPK5+akl2J#4PUmT
ztT#n%GteMNkjI$HL0{Nk$T?@ixc(<B92ZBO`?raMl_^g2>`vm7PE#ixt(t|eb+1rL
zseOH>DX$m#vh*g-&87iphq3FQI2{trHU26k_Ln*&3bHnIf<$C_PrB|}AN+!ukX{bw
zawv7*3>=~S7AU>7qF~k==MI0JuOQ$c!q5;s6-L#!qf^F?ux#xUePv})G!fkoqCZ`O
zF{|Mdtd3F-30Hc9Yx_5=^KxJvZESrb<>=Fs1dJ|xYCd6pWC8yShD9dDs8Ab#q&hj>
zorl*!&3&eLEkt-i?>=yNsmbiM{3P*oG^%ZM`E>iO-f;&OFWC2PMTm-us`cr#l$Gtb
zQLiWDBrxvm6Vx^jdbMp(6jw|e_(JS}k-8~NTuH4NafxwXIM7j!A(2~ln}Dh&1}6R&
zFHaPC)o-aQ>@H5HA{IWThdE8EhD`-}dxeel9otI@kr}PgF$rY|YAg?)>`gzle$~6J
zz1$xZl0I(7wSP%J-pd52s;P-UPW4Zg)rBhkrcqbwjYO3uQPyRkpCgZt&r$HdxEy?r
zoaI=nX>&r5{ipVYqbVDPUNT_^c#LWKsLPzOd>mgQ<#VvA@Xc?qc7DESbCWFq?NRva
zF;4~C5EME5V;jb2HQ@gwigRa&O<W8hui3$i8m9C{>}nNw+NIA7=!ZCH!c;pe{O|op
ztNi~Hy$J;WIEfd7U{*jU8Ps2;V1@A65$k6eCJ0?oD?`_yupbbfUqJ@@dGVSghOQK3
zDfE7R$Q~2cMbM&ChUO*m*?}VY3w7fy>$r-t>N0Pvtz$eQ=~p^tL@4;JWA{vmz88#E
z{)EEtH{3j~{|h<VM_)+%4=>=~aJfF>9)$s-O-TsFr^=85DxF=X74<l{ZqU#h>f90y
z$VikJG0E@lV(`loXRnI?2RsIL^z<|y^D1eGd|SNUfhL<?$};v3R35sjoz66Y0kiP9
zLZlP*l=aDz=?hSa+fkcfqXVmv?HhzB`xvTFBb72g1m~TDpv(T9#P`gN-4De3LmpMn
zK+vbmoo!=<v?N_MwQ9&*3+_I!2#<)+u3UGG8@wFlJ~}!|g%>GGyjRl=70X#FFfGid
z#MrJ$)c4)JnFFN8#N6I*-Qi++_txr(2(s91ChE6kCUU>WaC*1|{0t7p04lLu&@p>t
z0B+L9T~GOhqwPdQ0zTn4*ev-xo!pJ4+`Kd1=|*J<xLteRy$Hh%qfu=8NMMoT`FIfg
z_ptCM;cz7{8SeI{uM~ULlE^zAa~pBETmU!t#Y~U;F`JgCG-gI$2wi4r2)Y|^$b%v?
zo1RbVea=_%fJM9yypGA!#`WG!o?hxU6vA^g2A#x8CGD3(-wh4PzLnaHrNP~J21;Ol
z0<*~`mIox}C*X)0ZtP4?XG5;DEbv<*mN>}bi_d+>&XMaeud}Dz<AmYh#RPNWXABUZ
z6Skx2Kt~do(Dg*XnVN!$PBS;C52it}(Y<N&MR9Kbx?sVB7uv<3!_U4jPPgH7O1)HY
zvGyR0d-YyFVJ9|jC6|bbN<+b_qSPvS7=i}VvaLtq6^o5~T|0>TJ{q&rsHABVr!8xw
zyvS=WwrdWlE6K4$lG>s;k6@hA8zFrg_qVATXqT_K2tW<u+{s$^3)!PNJA4k>jRiAO
z1O$X2JA5P!W;LTa?OvmlPe4zQn}e3FTRyk$+B{L7a@|w80$`^j$UDA1A?>K)R51kP
zZFtdTMyeH!hAFHBmsNp$H!k$nA4fCZcDuO7gkvQkVJR;!U?HzE{)HagK02#W{wZ{V
zX`rPM39%I;<~{uz&U+eEVm_aX@iX7oGnQFCqDV*qqAr30gE{R{XY%)C@%n2QtRWaw
z$!{bUTKJio8WENQQfm|4aT99_qEsIk3_j2ZQg0eLnCvkyU{@z*=)`kzpqskOFj#qX
zVNUx~$$snb9`ibTK`3w|0X)CQ%U!Owu@e<lm6u-;MP(WpAwr1w{5j@m)`ko|TW@p4
z;;`A2V#KdoL|J<0HP?(!YpNHsrq{Kg63?`6qt9{qFv1J|&`)pevXAi1{j^E4{pFT@
zVDzkA=(!h{0rX@%!u|GoC-ibt(H`9QqhA4VyMPGMU8)?ig{Wj?b)gm#=k82T^VQ6^
zTV4lAGSsVjF}!^Z`fZgE1;=64f3crF4Hs7lgyv6}1V|<)V_aM!T=`HSYtfevM*Rqb
zUwM3g`(oyG4|>)K#)~I`yTA$Gbe1&rUa#6^ffEtb9L$EZPwA-b7hDYJBKrko9L(8L
z*0!H9fN0H-u7}1OqR2W`3|@!|;oD1f`O`U@1b8C;ccHt%#}H>dV*RLt&muVK#52}8
zz^RhqrcH1`j|G3jT*N@eKo$BaSahG2q7Ool9=Oj&=EX7xq(O6#&Kft5a(0wpN<FjI
zr~rMWVeNGZG?5qNNZtSK;`m?m$HTswD@Gn#kZ&XbqQM}WZsDiFa7ZIoq&y7!#DHHM
zTE~%`Dtpf10`$bCieZ;w%7u5FNB)3`34zY95Y@GqN`%nAU0^H>ip0OT4_mEmecoi;
zU$C)z`74@;PtafLE9d}M{Rj0uV)0P^RHQU;`M3UxY{ljj#dO1hdTXP>aDC#Ifgc7U
z;KpgJw$VN_?Q7lm@bV_NKdsb!#2qumlWwsaO^$0fpX7(!Q~{BG=KOHu+3TualNryz
z12c_`&%C9TMP&waF;?xIR-U&5hSC`(kc8@NAX)cG5$%0>i=n%e(rezx<ze&TBK4o~
z-_Muq4-auEDJg~W!#>+F;!}kz)-5F46WP*`2dOkA#dCbd8C~RePGO;^J2@L3)7!K6
zFCFSpDcKATS1}ixAIj1uiQLaG-!m`y6(5_TqqACDK}+=~uPbM-EVQ)tE6s|ezV4Gx
zX8@69?ArU21^^2N2%?c1E>z5r=wWwsME!{&>yJ*otE8?dI|oW=RKVqwB3(sUabXd|
z@2!97A(}mP&Cc-5QmG|j_T;H^5))UW&7uRVi&R&%@MN^-+$lQ+r^iT9wgL?<4#xd1
z+%V)cZ30h|gZJ7WqD&m|uX+NV6w~MkV%P>OYwXs%j^_`;^4>Hn*HPemH1s?e7#Sfr
zJDPlzrDu2UeZ&IrK!4oD#-6zzW!v>r656>@L%kk4(`VJ4v%WXl`Bpl5(V_b=K^rG?
zmcd!9Y<6!g?!PhyTU4IbU%P&bfXkxM`gTB&zRJPJQuy*i6^kMUFMRgsRFxhK4O8O$
zh+|%>c`~;WT|kT12+YdX#?7G+O)OtTROWMXq(0DqMS^$xqD4!u@$1ois(@wHeG8zK
z5O^yc*B431{pi;26=!l0+Y5igR<Xs#{Dr7Fc+GlVU45z6=tZH3sE)wJTaL#WN9cRi
zgQM0BZfU*!GT05P$tfa6vuyUwyKdt}QNxR%F2(QjfhbJ`;lpwFm0_vs9*e$PW+k<8
z4<x@4J~?n~f?U6rJy<V4n#y_umSZM0La9g82x=vAbi#;&-Y|XbcHiMFVT5tPrx5Lv
zx69%2(^VgX@r=N~b-wj9y!#k&`ELeq!<0=%0MC}X(peXjGTfWWq@mEK2MN?(dq!|^
zSf&X_ju3A}5<C|B&5(n=iX=2_msh5(@dSA`7JXu8>C#=JR&BfAf!Ai%Y-U?nl&6iH
z_2-a>6#oEfYa3fu|H}N~VKYSO6*AJ|VqyqzajC#;uz?4?xS11{fm!UYBkb+od+AY-
z$;RzYoGAZ@ldz@ta%85#jY1mx$t0<s1^bf0Y_R(s`7(KT03;l1@tWNhYJcM8;M%_i
zyt&_upE@n?8=jCnE+X3z{JR|JRW4X7>7>f-B4GA?4HF>?mIzUX^#6H=&81&#Rujlj
z2BgHqgiHldzup#mslSHCUJ~pRuYI;*yw2VDK7kv=2}WFyRb&ouF$gNGtxlk6)p(+8
zi2Yk7o?jtVLOK`0{rLiJ+i{`Z#w*if7i=+w<7o@=H#j+O8rFPCA-aiCQ7Rj^&}=!C
z!B-GDDnt4~G!znxbK$MbaV|H$7(kxx>y5y5*(A7NoTM&6RZS%JE&ISthnqfu)zc~5
zm!oC!-)%b)+^-kQPEo#3CuY&9YXCsw3S9zgk1(3ZZy65Cr_VLnq3)=ml1Y`eZezJ3
zC=vCU{C-l<K29NH%x6~x*=pj4vUjd+3Xr}C7Z;l)ghTjjLZ(lyY0*f)t%jq*VtwxS
zOD2$|j#ese5}5epDnUhC8?tX74BsBu|72X={(skkZd!ZHzt+JO^L_7x3iyu@QU!Y?
zgi`-PuXHQKZiH#aOQ`FJfGnd4iCX?^bLYQI$^WJ2^IyYL(_RrFrac^@%0@DbKfN{y
zar>KjT84Y8(*pO1AX9w>Lg>N33yQS3ntdV^_C3}sW_=++T`^g632Am;qj~nNJ?A9I
z?E4NO<qu6NR(8eJgeB#_M}nj%$#uJ&6lP0;2^nS8p|%g<OPl;Z0h>|#OcVplIOcGx
z>v1&4hEsWby-j6g*cdZ;RU5tTI!9C6DFlH9!7ECGyy2nil%+73nAx^}ji$8Q%(rLq
z&>V?8B<qO-pLWU(Kl@+WJ7%niFX2tDurz)4X~U*xiQP^ZeQOpaZb?VoU9}Z$k85E#
zqhVoT-{`UJ1Ajthb6y^0I={!PG22s867wudHbX+OE>9##$;}{-9<ol4j~Pi3tZyvS
zwmWxGVx(bLCod3sD)4h2ENiXeyG!S>KZaQG9_)_6s35zBF)z{z%jr7!g{2eKnmdv&
zE6e29P-ouno%mSwEUb$C%CF7<Z#tHT6IOE&9~H0z)@RXsCU>6l`Stlbq^SacHwE<l
zA`Jgfw&vD1XZO-(Fh{28zUj|PGp@VpbdGhg8ycId5;ukmzLC@E;^!GT>F5f_kE=)p
zlUP{LRd)47@P&-?v?w!vw0}fP#G=ZvNNHqoJ`Kpp_~iWYNJd9OGH4_{ru0@LX`~)`
z=?+(qK3Jkj?lJ6Ps!N0eg~o0Fz6s*A3U4_@t-4e-de5&z)6xu=9zufGj;7_{G!bT>
z!)y;>fBJ#qiaCIW?h<swR!z`exPl3ugHzbS)J`hD2_sgND6Dvxn3}Zss75KoQ=Gfy
zkch(L_t1IlALg3o<iNZ$1GiXn1rK`fs16AbcIBq?Xw(`my*EctdwSa(vq*Ud?(P)n
zBH{JeXi!}Fcyc$A?ATW1+}qGHfGnYG)I$8K{2Wimn_hvyV7=V9a`ypqcbQ6t=_0!^
zJQC?=B#ea*Oi5X1jFPkL)gy;Xa(t7WuwGc_^~2CFa2>MS%ID8tci>J?SJNvlHGG>)
zpH@W+`^@rD2#-%cF>H?v7zQG`l4#dCjQhc#q~U8Ak@IsD<ut!u+0|E;9pPRTV!9<f
z$d8a*yu?iDmo0x-_`+aE#GybeL{G(w(`75)xjO$oo1NOMP9P#aukjg_J7>MPH8|u-
z9Yx5?LFX#pd6()&Ld3~WM=>Li$i~M4_hJ7mvrN$Ouq#|z3u1+DA4N>G*_=)OYKe*^
zjLOMBbqaF#UGKObn*(B(r*oWr7&8ojqtb(zE&dKr>kVIhlbov;(v!gO9=f`H5<iv6
zCn5a=6MTW$a<mh%7`ahtv`7dj&(05~MeoBx88ClPKm<BNP#@`tluy_BdRkl){!dZ0
z1j@xRv}MOk>TO3?EOu(<+l=mFmR}I{3+bw!h<538HKfNZ$zD@II#S3x=N#ggYtC{K
z#v>-1T{Z59Z1)t|wXL|ksw}u<8ZS=x(~M(7+I(VW4djVa-0zLapatYKg+KOiFB^2K
zLRx@0o&-NJ#PfMW?;3FTn9M=L6>ns`WFH~7GYaLW8B(IaMf&G_PnmPThs5zUf0*aM
zKL!ohf`iR#JN+wlRM)mG%B&`kWkN0)8#qO{gt0smTpk&I{FI1;am6l}`C9S3>M@m-
zdM3?*3dnD**W1%-Yi{;013oR1-g+~E+l{s>DH}c7fSd5AbG8dL_On=co3(sttXj2a
z1NeVxsE}s--o1l<{`-!AL2Qa?JoBvU#pcesfjo^>7h)dYM7Tb#X&*ia&sQwt632Id
zaadSK<}>po)A&&%5Z~;tA%%cViWSssddE>qJ+Y=*0aK#R6vVKH(5abi_FjViM2N`l
z)mMYL;iULJ<(f&!T|yk_-D-{L{HJioG_PL4^}7w1d{*;`e^%Unava^SqMuOruPv8v
zk7w8#QKt3&0sHG0PA#Zu&CTc;OW)tLQAFV&fRrRimV2&}OjGAWNOB!8+$-YC%har1
z%4KY3=3Gr(woWJJv#nr;E87ZP1uvwxFWTt7YUxz;j5T3?mT6{7HXYGQ(!y9=6y)9X
zV^K3%4KX)DiB*rW+?hu!aFUIN1rDXg-ysI%fS&O{8qMoyFQ%zfDUr@cU3VCah;FL;
zi4@y$A-zG-uQd=uh1?a68!J^i0<sPVmsA$zS&kHuRq`62#-Wa-C#0TQ1UDTkCfLOe
zCc(JLHJ-*XGFOmrd%72KO(@cNIZvKAU9J2mh*}eZ%c$=(_D68uY#T~y@eGiJ^zPs^
zD7>zkGg{^5)8ip8|9<Ld^9lZ<ZL>G_yE34NPbDp(ctaSJNU*$T0D`z(obO**`!H}Y
zjDMp1oI08p->2SkaY3lrVBkK^;PlaXJQt4UBy=MJ9%3v)x>hf;f3oXqdUHFMkK=;R
zggj?!d`(=syYH4^sq$70^xA9-3(;k!K|dsS6^D$qTo)FJnQ;O4+Yidcff<lXJKtKa
zd{*I%s#?BO&7z5UzMsPv((-0IU-{CZuGrlS>3kKxDu|TB-^X93Je_P~*z&RBI(gX<
zU?**{<6;zzWYORAEgi`aEi`Qqv~dp+?H})c=Hnm{@F1gL!nd5K(zm1L#yfKi#Y<Kb
z=qO+TJe>gY3k#+5<yu@uOovnLUUOZ<M7Dnz$;?cOiMed7>?_cG*Hed^ayah}HnZXQ
zz8EBi)(i#(84H<msiB`ioHKFB7KweG24pY)#LOnQlLQ|#MA<A-o0<lF{KvTyg=nDX
z)uF#KD(wrc%=FJg3C04$CD*eze@{$|_LY=Mz}ng`%4j)s>v9I?Mdu0mvGGiPLlIY*
zP%qfykz^-6x6`Fsi<T$$P};@*Myz3|ZQ(%cdul0QGe`w{;Rez&Fe<Z|q1H48kK40W
zrzKtP#@5v9nVymB<$U>ND^3V?mv(&?l7E(2`veJaBiqz)(2K^|FsUp3uw89+nkpd9
z(7<*35(+WrI(|oKA-GU&kp^jjl<p2aUFd*|`WE$z<)e(03HRt3Ccl`{x(9gWex0da
zi{<fAwM2t|k*@FVy}sivPAL?-e{T!#!jf4cOQzlY$V<R2Fl!Ouhsh9#|E3fgt5n-K
zEPSx>>TZA={=;6O5Nvi4P6})$_4WH+;Q$b!^1X)<d4ba_h5MWD{vS~4|7kdTIk64H
z(3~8o6f8#WrlwLhC83HDfF@D|g+=9Glvid-Ol1l$0-uedWRyI!EXI2R8~LL*T<D)S
ze?lZ9>-bJ2?uzo8@>ATZFp5(O^2v$%AusvNUdV)q!As!oaS&QcE%nmKG0lXlQK?$u
zNxu`{z@|eJY50<~<bNTvA1JDOVDaz<c~k1`EyCB$4zdCJR|lcqeb?y-*;C#Zhm5?F
M7FQ4}7cmI<Z?Jlm$N&HU

diff --git a/docs/contributing/principles-and-process.md b/docs/contributing/principles-and-process.md
index b0bebbdd..4dd21190 100644
--- a/docs/contributing/principles-and-process.md
+++ b/docs/contributing/principles-and-process.md
@@ -3,15 +3,8 @@ title: Principles & process
 ---
 
 <!-- The rules live in CLAUDE.md at the repo root, where every Claude Code session
-     auto-loads them and where GitHub renders them. This page embeds that file rather
-     than restating it, so the site and the working rulebook are the same text (one
-     home, per the no-duplication rule); pymdownx.snippets resolves the path from the
-     repo root via base_path.
+     auto-loads them and where GitHub renders them. This page embeds that file rather than restating it, so the site and the working rulebook are the same text (one home, per the no-duplication rule); pymdownx.snippets resolves the path from the repo root via base_path.
 
-     The `title:` above names the page in the nav and the browser tab. The embedded
-     text opens with its own `# CLAUDE.md` heading, which is the right name at the repo
-     root but not in a docs menu — front matter overrides it without editing the source
-     file. The doc links inside the embed are rebased at build time (docs/x.md → x.html);
-     see _rebase_repo_root_doc_links in moondeck/docs/mkdocs_hooks.py. -->
+     The `title:` above names the page in the nav and the browser tab. The embedded text opens with its own `# CLAUDE.md` heading, which is the right name at the repo root but not in a docs menu — front matter overrides it without editing the source file. The doc links inside the embed are rebased at build time (docs/x.md → x.html); see _rebase_repo_root_doc_links in moondeck/docs/mkdocs_hooks.py. -->
 
 --8<-- "CLAUDE.md"
diff --git a/docs/explanation/architecture/moonbase.md b/docs/explanation/architecture/moonbase.md
index 2f5529f6..842a7ff9 100644
--- a/docs/explanation/architecture/moonbase.md
+++ b/docs/explanation/architecture/moonbase.md
@@ -28,65 +28,28 @@ Neither image can rewrite the partition it is executing from, so each installs t
 
 ## What it replaces
 
-Dual-OTA spends half the app area on a second copy of the firmware that is idle except during an
-update. **MoonBase** replaces it: a small, rarely-changing image in the partition table's
-`factory` slot that owns the device while the application is being replaced, since a board
-cannot rewrite the partition it is executing from. One app slot then suffices, and the flash the
-second slot held goes elsewhere.
+Dual-OTA spends half the app area on a second copy of the firmware that is idle except during an update. **MoonBase** replaces it: a small, rarely-changing image in the partition table's `factory` slot that owns the device while the application is being replaced, since a board cannot rewrite the partition it is executing from. One app slot then suffices, and the flash the second slot held goes elsewhere.
 
-A 4 MB board has no choice, having room for one application and not two, and its app slot grows
-by a third in exchange. On a **16 MB** board the choice is deliberate rather than forced, and the
-freed 4 MB goes to the filesystem, 11 MB rather than 7.
+A 4 MB board has no choice, having room for one application and not two, and its app slot grows by a third in exchange. On a **16 MB** board the choice is deliberate rather than forced, and the freed 4 MB goes to the filesystem, 11 MB rather than 7.
 
-Which boards use MoonBase is a per-variant decision recorded in
-`moondeck/build/build_esp32.py` rather than a property of flash size. Today the 4 MB classic, the
-S3-Zero and `esp32-16mb` use it, and it may become the default everywhere.
+Which boards use MoonBase is a per-variant decision recorded in `moondeck/build/build_esp32.py` rather than a property of flash size. Today the 4 MB classic, the S3-Zero and `esp32-16mb` use it, and it may become the default everywhere.
 
 ## The update cycle
 
-The update cycle runs in three moves. The app stages the install URL in NVS, or nothing for a
-browser upload, points the bootloader at MoonBase and reboots. MoonBase joins the network with
-the app's stored credentials, falling back to an AP at 4.3.2.1, and installs into the single app
-slot, either from the staged URL unattended or from an upload. Then it reboots back. The UI covers the whole cycle with one
-"updating firmware" overlay, telling the two images apart via `GET /moonbase` (MoonBase answers
-with its live status; the app 404s it). Pointing the bootloader at a factory partition *erases*
-otadata, so a power cut anywhere mid-install boots MoonBase and the user retries over the
-network, a stronger power-fail story than dual-OTA's. A failed install deliberately stays in
-MoonBase, visibly, rather than silently reverting to the old app; the way back is its explicit
-"Boot the app" action, which only boots an image that validates.
+The update cycle runs in three moves. The app stages the install URL in NVS, or nothing for a browser upload, points the bootloader at MoonBase and reboots. MoonBase joins the network with the app's stored credentials, falling back to an AP at 4.3.2.1, and installs into the single app slot, either from the staged URL unattended or from an upload. Then it reboots back. The UI covers the whole cycle with one "updating firmware" overlay, telling the two images apart via `GET /moonbase` (MoonBase answers with its live status; the app 404s it). Pointing the bootloader at a factory partition *erases* otadata, so a power cut anywhere mid-install boots MoonBase and the user retries over the network, a stronger power-fail story than dual-OTA's. A failed install deliberately stays in MoonBase, visibly, rather than silently reverting to the old app; the way back is its explicit "Boot the app" action, which only boots an image that validates.
 
 ## Updating MoonBase itself
 
 **Updating MoonBase itself** runs the same cycle backwards: the app writes the factory slot while
-running from `ota_0`, exactly as MoonBase writes the app slot while running from factory. Neither
-image can rewrite the partition it executes from, so each installs the other and the app is the
-only thing that can repair a broken recovery image. Without it a bad MoonBase means a cable, which
-is the failure MoonBase exists to prevent.
-
-Two things make that safe enough to offer. `esp_ota_*` refuses a factory partition, so this is a
-raw `esp_partition_erase_range` + `esp_partition_write`, which also forfeits the validation
-`esp_ota_end` performs: `esp_image_verify` replaces it after the write. A 4 MB board also has nowhere to stage 743 KB before erasing, so the image streams straight in.
+running from `ota_0`, exactly as MoonBase writes the app slot while running from factory. Neither image can rewrite the partition it executes from, so each installs the other and the app is the only thing that can repair a broken recovery image. Without it a bad MoonBase means a cable, which is the failure MoonBase exists to prevent.
+
+Two things make that safe enough to offer. `esp_ota_*` refuses a factory partition, so this is a raw `esp_partition_erase_range` + `esp_partition_write`, which also forfeits the validation `esp_ota_end` performs: `esp_image_verify` replaces it after the write. A 4 MB board also has nowhere to stage 743 KB before erasing, so the image streams straight in.
 Everything that can reject it is therefore decided from its FIRST CHUNK, before a byte is erased:
 the image magic, the chip id, and the descriptor naming `projectMM-moonbase` rather than the app.
-The chip id matters because there is one MoonBase per chip, one paste apart, and a checksum does
-not catch a swap. Those rules live in `src/core/util/FirmwareImage.h` so
-a host test can drive them. What remains is a window, during the write, in which the device holds
-no recovery image; the app keeps running throughout, so the answer to a failure is to retry.
+The chip id matters because there is one MoonBase per chip, one paste apart, and a checksum does not catch a swap. Those rules live in `src/core/util/FirmwareImage.h` so a host test can drive them. What remains is a window, during the write, in which the device holds no recovery image; the app keeps running throughout, so the answer to a failure is to retry.
 
 ## Telling the two images apart
 
-Each image reports its version from the app descriptor IDF puts in every binary, with
-`PROJECT_VER` set to the same computed version for both. The app can therefore read the factory
-partition's version without booting it, and say when the two were built apart. A device that cannot name its
-own recovery image cannot be diagnosed: two boards that looked identical, one of which could not
-install firmware, took a bisect of the git log to tell apart.
-
-MoonBase is a standalone ESP-IDF project (`moonbase/`, ~750 KB against an 896 KB slot) sharing
-no sources with the app, the deliberate trade for an image that must stay small and, once
-working, hardly change. `moondeck/build/build_esp32.py` builds it alongside every variant that opts in, and owns the
-flash-layout helpers every consumer uses: serial flash, mooninstaller manifests, release preview
-and the QEMU image. IDF's own `flasher_args.json` knows nothing of the two-image scheme and
-stages the app at the factory offset, so each of those paths applies the same correction from one
-place. Prior art: Tasmota's safeboot scheme and
-[MycilaSafeBoot](https://github.com/mathieucarbou/MycilaSafeBoot) proved the single-slot +
-recovery-image pattern; MoonBase is our from-scratch, minimal take on it.
+Each image reports its version from the app descriptor IDF puts in every binary, with `PROJECT_VER` set to the same computed version for both. The app can therefore read the factory partition's version without booting it, and say when the two were built apart. A device that cannot name its own recovery image cannot be diagnosed: two boards that looked identical, one of which could not install firmware, took a bisect of the git log to tell apart.
+
+MoonBase is a standalone ESP-IDF project (`moonbase/`, ~750 KB against an 896 KB slot) sharing no sources with the app, the deliberate trade for an image that must stay small and, once working, hardly change. `moondeck/build/build_esp32.py` builds it alongside every variant that opts in, and owns the flash-layout helpers every consumer uses: serial flash, mooninstaller manifests, release preview and the QEMU image. IDF's own `flasher_args.json` knows nothing of the two-image scheme and stages the app at the factory offset, so each of those paths applies the same correction from one place. Prior art: Tasmota's safeboot scheme and [MycilaSafeBoot](https://github.com/mathieucarbou/MycilaSafeBoot) proved the single-slot + recovery-image pattern; MoonBase is our from-scratch, minimal take on it.
diff --git a/docs/explanation/architecture/mooncore.md b/docs/explanation/architecture/mooncore.md
index 1ada3a18..03b7a489 100644
--- a/docs/explanation/architecture/mooncore.md
+++ b/docs/explanation/architecture/mooncore.md
@@ -111,14 +111,6 @@ Abstractions are added when a concrete implementation needs them, not pre-design
 **Platform boundary (hard rule).** All `#ifdef`, `#if defined`, platform-specific `#include`s, and hardware API calls live exclusively in `src/platform/`. Everything outside `src/platform/` compiles on every target without modification. Compile-time platform branching uses `if constexpr` on `platform_config.h` flags, never a preprocessor `#ifdef`. The boundary is enforced by [`moondeck/check/check_platform_boundary.py`](../../moondeck/check/check_platform_boundary.py), a commit gate (see [CLAUDE.md § The Process](../../CLAUDE.md#the-process)).
 
 **The desktop build runs everything (hard rule).** Every module, effect and driver in the repo
-links and runs on the host, the platform layer has no silicon behind the call. Where a
-peripheral is absent the host *emulates* it rather than declaring itself incapable: the parallel
-WS2812 buses are backed by heap buffers, `lcdLanes` / `parlioLanes` / `rmtTxChannels` report a
-real chip's counts, and `hasLcdCam` is true. Code excluded from the host binary is code that cannot be unit-tested and cannot be seen by any
-AST-based check. It only ever runs where it is hardest to debug, which is what the LED drivers
-were until they were linked here.
-
-A capability flag therefore answers *"can this build exercise the path?"*, not *"is this real
-hardware?"*. Where a flag must mean the latter (`hasLcdCam` gating the pin expander), that is a
-deliberate, commented exception. Timing, wire protocol and pin state are NOT emulated: they need
-silicon, and faking them would let a self-test report on hardware it never touched.
+links and runs on the host, the platform layer has no silicon behind the call. Where a peripheral is absent the host *emulates* it rather than declaring itself incapable: the parallel WS2812 buses are backed by heap buffers, `lcdLanes` / `parlioLanes` / `rmtTxChannels` report a real chip's counts, and `hasLcdCam` is true. Code excluded from the host binary is code that cannot be unit-tested and cannot be seen by any AST-based check. It only ever runs where it is hardest to debug, which is what the LED drivers were until they were linked here.
+
+A capability flag therefore answers *"can this build exercise the path?"*, not *"is this real hardware?"*. Where a flag must mean the latter (`hasLcdCam` gating the pin expander), that is a deliberate, commented exception. Timing, wire protocol and pin state are NOT emulated: they need silicon, and faking them would let a self-test report on hardware it never touched.
diff --git a/docs/friend-repos/FastLED-FastLED.md b/docs/friend-repos/FastLED-FastLED.md
index ceb205ff..b8ceba31 100644
--- a/docs/friend-repos/FastLED-FastLED.md
+++ b/docs/friend-repos/FastLED-FastLED.md
@@ -1,6 +1,6 @@
-# FastLED — monthly activity digest
+# FastLED: monthly activity digest
 
-What landed on [FastLED](https://github.com/FastLED/FastLED)'s main branch, month by month. External-context reference (like the v1/v2/MoonLight inventories) — a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these digests lives in [README.md](index.md).
+What landed on [FastLED](https://github.com/FastLED/FastLED)'s main branch, month by month. External-context reference (like the v1/v2/MoonLight inventories), a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these digests lives in [README.md](index.md).
 
 ## August 2026
 
@@ -47,10 +47,10 @@ _Auditability: 195 first-parent commits on `master` with author-date 2026-08-01.
 No release cut this month (3.10.4, 2026-06-16, remains the latest), so the month is not split. Two big threads: finishing the Raspberry Pi Pico driver family, and cutting the ESP32 platform loose from the Arduino core.
 
 **New**
-- Raspberry Pi Pico / RP2040: automatic parallel PIO output finally works for real — 2/4/8 strips driven from one PIO program, with a single-lane fallback for mixed layouts.
+- Raspberry Pi Pico / RP2040: automatic parallel PIO output finally works for real, 2/4/8 strips driven from one PIO program, with a single-lane fallback for mixed layouts.
 - Raspberry Pi Pico gains fixed-function SPI+DMA drivers, a UART DMA driver, and a public hardware-SPI routing API.
 - WS2814 RGBW strips are now a first-class chipset with datasheet timing.
-- Classic ESP32 gains a second I2S bank — up to 32 parallel strip outputs — plus a second UART output lane.
+- Classic ESP32 gains a second I2S bank, up to 32 parallel strip outputs, plus a second UART output lane.
 - FastLED can be built as a plain ESP-IDF project with no Arduino core at all: IDF's own time, serial, SPI, LEDC and heap calls are now the default on ESP32, with Arduino only as an opt-in fallback.
 - Classic ESP32 also gains an I2S-based signal capture backend (reading WS2812 data in), alongside the existing RMT and LPC845 capture paths.
 - LPC845 now defaults to its UART DMA output path.
@@ -69,10 +69,10 @@ No release cut this month (3.10.4, 2026-06-16, remains the latest), so the month
 
 **Watching**
 - Report that RGBW output has been broken since 3.10.3 (#3622, closed) fed the month's RGBW colorimetry cleanups.
-- An open thread (#3762) blames an unconditional deep yield in `show()`'s refresh throttle for a long-standing frame-timing regression — no fix shipped yet.
+- An open thread (#3762) blames an unconditional deep yield in `show()`'s refresh throttle for a long-standing frame-timing regression, no fix shipped yet.
 - A port to the WCH CH32V003 (48 MHz, 2 KB RAM) is proposed (#3755).
 
-_Auditability: 212 first-parent commits on `master` with author-date 2026-07-01..2026-07-31. Issues via `search/issues` for `repo:FastLED/FastLED+is:issue+created:2026-07-01..2026-07-31` (110 opened) and `closed:2026-07-01..2026-07-31` (106 closed); the great majority are the project's own phase/meta bring-up trackers for RP2040, LPC845 and the classic-ESP32 I2S driver, plus CI and linter work — only the user-facing ones are surfaced above. No versioned release published in July, so no month split._
+_Auditability: 212 first-parent commits on `master` with author-date 2026-07-01..2026-07-31. Issues via `search/issues` for `repo:FastLED/FastLED+is:issue+created:2026-07-01..2026-07-31` (110 opened) and `closed:2026-07-01..2026-07-31` (106 closed); the great majority are the project's own phase/meta bring-up trackers for RP2040, LPC845 and the classic-ESP32 I2S driver, plus CI and linter work, only the user-facing ones are surfaced above. No versioned release published in July, so no month split._
 
 ## June 2026 (up to 3.10.4)
 
@@ -80,7 +80,7 @@ Released **3.10.4** (2026-06-16), cut from `master`.
 
 **New**
 - STM32: Arduino UNO Q board support.
-- New NXP LPC8xx family drivers land (LPC804 PLU, LPC845 bit-bang + PWM/DMA-to-GPIO, LPC11xx) — early bring-up, bench-validated.
+- New NXP LPC8xx family drivers land (LPC804 PLU, LPC845 bit-bang + PWM/DMA-to-GPIO, LPC11xx), early bring-up, bench-validated.
 - Unified `fl::Watchdog` API with real hardware implementations across platforms (ESP32, Teensy 4, AVR, Apollo3, RP2040, STM32) plus a non-allocating reset/crash-classification helper.
 - Wave simulation: opt-in 9-point isotropic Laplacian (smoother 2D waves), exposed as a UI toggle in several example sketches.
 - ScreenMap gains a v2 schema (auto-detected), and a new `.fled` container format for video/screenmaps that `FxSdCard` can load.
@@ -114,13 +114,13 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 **New**
 
-- New **Channels API** for managing multiple LED drivers at once — a `fl::Bus` type, `FastLED.add<Bus>(...)`, `fl::enableAllDrivers()`, and `FastLED.setExclusiveDriver(...)`, with a diagnostic that warns when a strip's driver doesn't match its bus. (The month's biggest effort.)
+- New **Channels API** for managing multiple LED drivers at once, a `fl::Bus` type, `FastLED.add<Bus>(...)`, `fl::enableAllDrivers()`, and `FastLED.setExclusiveDriver(...)`, with a diagnostic that warns when a strip's driver doesn't match its bus. (The month's biggest effort.)
 - **RGBW / RGBWW color**: proper colorimetric RGB→RGBW conversion with a lookup table, color-temperature (CCT) control, and an RGB+CCT mode.
 - ESP32-P4 gains a SIMD (PIE) acceleration backend for faster pixel processing.
 
 **Faster**
 
-- Big speedups to the ESP32-P4 **PARLIO** parallel driver — encoding and transmission now overlap, and a chipset-aware encode path is ~5× faster than before.
+- Big speedups to the ESP32-P4 **PARLIO** parallel driver, encoding and transmission now overlap, and a chipset-aware encode path is ~5× faster than before.
 - ESP32-P4 "Wave8" output ~1.2–2× faster via new transpose and lookup-table paths.
 
 **Hardware & build**
@@ -140,7 +140,7 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 **New**
 
-- **Audio "silence gate"** across the audio-reactive features — tempo, spectral metrics, and the Vibe effect now fade out cleanly when the input goes quiet instead of reacting to noise.
+- **Audio "silence gate"** across the audio-reactive features, tempo, spectral metrics, and the Vibe effect now fade out cleanly when the input goes quiet instead of reacting to noise.
 - Audio FFT can run on the ESP-DSP hardware backend (faster spectrum analysis on ESP32).
 - ESP32-S3 LCD driver gains ISR-driven chunked DMA output (smoother large-strip output); coroutine tasks can be pinned to a chosen core.
 - RMT receive (reading signals in) gains DMA streaming; a long-strip SPI bug (#2254) fixed.
@@ -154,7 +154,7 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 *Summarised from 444 first-parent commits on `master`, 2026-03-01 … 2026-03-31.*
 
-- Mostly an internal stability and build-correctness month (sanitizer fixes, WASM build speed, IWYU/PCH hygiene) — little user-facing.
+- Mostly an internal stability and build-correctness month (sanitizer fixes, WASM build speed, IWYU/PCH hygiene), little user-facing.
 - **Fixed:** AVR builds (replaced defaulted `noexcept` with explicit implementations in container types); ESP32-C3/C5 and Teensy build breakages; printf/Arduino-compatibility shim; an audio-path bug.
 
 ## February 2026
@@ -168,7 +168,7 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 *Summarised from 582 first-parent commits on `master`, 2026-01-01 … 2026-01-31.*
 
-- A heavy **stability-hardening month** — most of the work was fixing memory and initialization bugs surfaced by sanitizers (ASan/LSan/UBSan): use-after-free, memory leaks, static-initialization-order issues, shared-pointer errors.
+- A heavy **stability-hardening month**, most of the work was fixing memory and initialization bugs surfaced by sanitizers (ASan/LSan/UBSan): use-after-free, memory leaks, static-initialization-order issues, shared-pointer errors.
 - **Fixed (user-visible):** a crash in power management; RP2350 system defines not being included; an ISR error where an int was read as a bool; i2s/LCD-CAM.
 
 ## December 2025
@@ -177,7 +177,7 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 **New**
 
-- **PARLIO driver maturation** (ESP32 parallel output): streaming support, up to 16 lanes, larger per-channel LED counts, background DMA buffer worker, and a low-level hardware abstraction layer — plus many alignment/timing fixes (the "1-bit shift" buffer-boundary bug).
+- **PARLIO driver maturation** (ESP32 parallel output): streaming support, up to 16 lanes, larger per-channel LED counts, background DMA buffer worker, and a low-level hardware abstraction layer, plus many alignment/timing fixes (the "1-bit shift" buffer-boundary bug).
 - **Validation / proof-of-life framework**: hardware-in-the-loop validation, an ESP32 watchdog (with a USB-disconnect fix), and a result banner.
 - New `Potentiometer` class (hysteresis + calibration); 16-bit PWM pin support (`setPwm16`); per-channel gain on the HD108 chipset; `Serial` gains `printf`.
 - Signal **receive (RX)** gains raw edge-time capture and a safe sketch-halt.
@@ -188,10 +188,10 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 **New**
 
-- **Channels / ChannelBusManager foundation** — a unified, priority-based driver manager with fallback, centralized SPI driver registration, and a new ChannelEngine-based SPI driver + RMT4 driver for ESP32 IDF 4.x. (Start of the multi-driver architecture that continues through to May.)
+- **Channels / ChannelBusManager foundation**: a unified, priority-based driver manager with fallback, centralized SPI driver registration, and a new ChannelEngine-based SPI driver + RMT4 driver for ESP32 IDF 4.x. (Start of the multi-driver architecture that continues through to May.)
 - **PARLIO** gains a runtime-configurable multi-channel driver with auto-select (and dropped ESP32-S3, which uses LCD instead).
 - New **UCS7604** controller; a generic clockless waveform generator; video playback support.
-- **Audio-reactive** effects expand — downbeat-darkness effect, AnimartrixRing audio reactivity, configurable UI audio.
+- **Audio-reactive** effects expand, downbeat-darkness effect, AnimartrixRing audio reactivity, configurable UI audio.
 - WASM web preview moves to dedicated worker threads (drops Asyncify) with incremental/PCH build speedups.
 - Experimental RISC-V interrupt support.
 
@@ -214,9 +214,9 @@ _Auditability: 481 commits with author-date in 2026-06-01..2026-06-30 on `master
 
 **New**
 
-- **WASM web-preview overhaul** — Three.js-based tile rendering, instanced LED rendering, SharedArrayBuffer zero-copy frames, a background-worker async controller, and an improved video recorder (native `captureStream`, 60 FPS, better MP4 compatibility).
+- **WASM web-preview overhaul**: Three.js-based tile rendering, instanced LED rendering, SharedArrayBuffer zero-copy frames, a background-worker async controller, and an improved video recorder (native `captureStream`, 60 FPS, better MP4 compatibility).
 - New hardware-accelerated **ezWS2812** GPIO + SPI drivers for Silicon Labs MGM240 / EFR32MG24 (MG24) boards.
-- **Codec support** — progressive JPEG decoding (4 ms time budget), and metadata parsing for GIF/JPEG/MPEG1.
+- **Codec support**: progressive JPEG decoding (4 ms time budget), and metadata parsing for GIF/JPEG/MPEG1.
 - Bilinear interpolation for upscaling effects; `FxNoiseRing` low-memory mode.
 - README/wiring guidance for high-parallel LED setups (incl. ObjectFLED parallel capacity).
 
diff --git a/docs/friend-repos/MoonModules-WLED-MM.md b/docs/friend-repos/MoonModules-WLED-MM.md
index ab14d3fa..d646b689 100644
--- a/docs/friend-repos/MoonModules-WLED-MM.md
+++ b/docs/friend-repos/MoonModules-WLED-MM.md
@@ -1,6 +1,6 @@
-# WLED-MM — monthly activity digest
+# WLED-MM: monthly activity digest
 
-What landed on [WLED-MM](https://github.com/MoonModules/WLED-MM)'s `mdev` (default) branch, month by month. External-context reference — a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md). Months are split at versioned-release boundaries (the rolling `nightly` tag is not a release).
+What landed on [WLED-MM](https://github.com/MoonModules/WLED-MM)'s `mdev` (default) branch, month by month. External-context reference, a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md). Months are split at versioned-release boundaries (the rolling `nightly` tag is not a release).
 
 ## August 2026
 
@@ -14,11 +14,11 @@ _Auditability: 1 commit on `mdev`, author-date 2026-08-01..2026-08-31 (f2d32c9c,
 
 *Summarised from 2 commits on `mdev`, both 2026-07-01 (no versioned release cut this month; the `nightly` prerelease republished on 2026-07-02 packages June's work).*
 
-Near-dormant month — a single small change and a build-number bump.
+Near-dormant month, a single small change and a build-number bump.
 
 - The instance list now shows which repo (WLED-MM or upstream WLED) each discovered instance runs, and reports the right release for upstream WLED instances.
 
-*Auditability: 2 commits on `mdev`, author-date 2026-07-01..2026-07-31 (range 12b0238 … 7c55f91; 7c55f91 is a build-number bump, omitted as not user-facing). Issues checked: `repo:MoonModules/WLED-MM is:issue created:2026-07-01..2026-07-31` (0) and `closed:2026-07-01..2026-07-31` (0), and `updated:2026-07-01..2026-07-31` (0) — no issues opened, closed, or even commented on all month. (The June window returns 2 on the same query form, so the zeros are real rather than a mistyped repo name: the search API needs `MoonModules/WLED-MM`, not `MoonModules/WLED`.)*
+*Auditability: 2 commits on `mdev`, author-date 2026-07-01..2026-07-31 (range 12b0238 … 7c55f91; 7c55f91 is a build-number bump, omitted as not user-facing). Issues checked: `repo:MoonModules/WLED-MM is:issue created:2026-07-01..2026-07-31` (0) and `closed:2026-07-01..2026-07-31` (0), and `updated:2026-07-01..2026-07-31` (0), no issues opened, closed, or even commented on all month. (The June window returns 2 on the same query form, so the zeros are real rather than a mistyped repo name: the search API needs `MoonModules/WLED-MM`, not `MoonModules/WLED`.)*
 
 ## June 2026
 
@@ -26,7 +26,7 @@ Near-dormant month — a single small change and a build-number bump.
 
 **New**
 - Waveshare ESP32-S3 Matrix Driver board profile added.
-- Audio-reactive is now an out-of-tree usermod (pulled in as a dependency) rather than baked into the tree — no user-facing change to how it works, but a cleaner build.
+- Audio-reactive is now an out-of-tree usermod (pulled in as a dependency) rather than baked into the tree, no user-facing change to how it works, but a cleaner build.
 
 **Fixed**
 - Output settings no longer revert to defaults after a reboot (regression in 14.7.2 nightly, reported on ESP32-D0WDQ5 and ESP32-S3; issue #367).
@@ -36,13 +36,13 @@ Near-dormant month — a single small change and a build-number bump.
 - ArtiFX (ARTI effect engine) robustness: call-stack bounds checks, safer string/lexer handling, and fixes for glitches seen only in speed-optimised builds (relates to issue #295); ARTI status now shown in the Info panel.
 - Upstream compatibility: accepts `I2CSDAPIN`/`I2CSCLPIN` as alternatives to the older I2C pin defines; fixed the arduinoFFT dependency; brown-out restart info now checked on both cores.
 
-*Auditability: 38 commits on `mdev`, author-date 2026-06-01..2026-06-30 (range 84669c3 … 70fe1b8; several are CHANGELOG/version-bump/CodeRabbit-config/internal-refactor commits, omitted as not user-facing). Issues checked: created 2026-06-01..2026-06-30 and closed 2026-06-01..2026-06-30 — 4 relevant surfaced (#367, #371 fixed this month; #105, #295 long-standing, closed/addressed this month).*
+*Auditability: 38 commits on `mdev`, author-date 2026-06-01..2026-06-30 (range 84669c3 … 70fe1b8; several are CHANGELOG/version-bump/CodeRabbit-config/internal-refactor commits, omitted as not user-facing). Issues checked: created 2026-06-01..2026-06-30 and closed 2026-06-01..2026-06-30, 4 relevant surfaced (#367, #371 fixed this month; #105, #295 long-standing, closed/addressed this month).*
 
 ## May 2026
 
 *Summarised from 34 first-parent commits on `mdev`, 2026-05-01 … 2026-05-30.*
 
-- **Ethernet board support:** added QuinLED v4 Ethernet profiles, a legacy Olimex ETH-Gateway option, and fixed the KIT-VE PHY address. **Breaking:** a duplicate Ethernet option was removed — re-select your board if you used the previous Olimex-ESP32-Gateway entry.
+- **Ethernet board support:** added QuinLED v4 Ethernet profiles, a legacy Olimex ETH-Gateway option, and fixed the KIT-VE PHY address. **Breaking:** a duplicate Ethernet option was removed, re-select your board if you used the previous Olimex-ESP32-Gateway entry.
 - Audio-reactive auto-disables during DDP / DMX / Art-Net input (avoids the two fighting over the LEDs).
 - Persistent on-screen error display when a restart is needed (errors no longer scroll away unseen); Improv and MQTT input hardened against malformed data.
 - Web UI accessibility improvements.
@@ -51,7 +51,7 @@ Near-dormant month — a single small change and a build-number bump.
 
 *Summarised from 47 first-parent commits on `mdev`, 2026-04-01 … 2026-04-30.*
 
-- **DDP input** hardened — rejects malformed / unsupported / "control" packets, relaxed header checks for compatibility.
+- **DDP input** hardened, rejects malformed / unsupported / "control" packets, relaxed header checks for compatibility.
 - Robustness: recovers gracefully from an empty `{}` config file (`cfg.json` / `wsec.json`) instead of misbehaving; steadier serial on ESP32; ESP8266 build fixes.
 - Otherwise a documentation / AI-contributor-guideline month (little user-facing).
 
@@ -73,7 +73,7 @@ Near-dormant month — a single small change and a build-number bump.
 
 *Summarised from 42 first-parent commits on `mdev`, 2026-02-01 … 2026-02-28.*
 
-- **Memory / "Heap too low" work:** moved the WS-LED preview buffer into PSRAM, PSRAM-aware allocation, reduced JSON buffers on S3-without-PSRAM — fewer out-of-memory failures on tight boards.
+- **Memory / "Heap too low" work:** moved the WS-LED preview buffer into PSRAM, PSRAM-aware allocation, reduced JSON buffers on S3-without-PSRAM, fewer out-of-memory failures on tight boards.
 - **Board support:** ESP32-S3 QSPI build, builds without the HUB75 driver (4 MB / 16 MB variants), better handling of ESP32 PICO-D2/V3 and D0WDR2-V3 (frees GPIO17), startup serial now prints HUB75 pins + full chip revision.
 - Audio: disabled broken I2S 16-bit sampling; fixed Ethernet errors when using I2S audio.
 - Spots effect fixes; fixed a short black-out when a playlist advances.
@@ -82,14 +82,14 @@ Near-dormant month — a single small change and a build-number bump.
 
 *Summarised from 48 first-parent commits on `mdev`, 2026-01-13 (after v14.7.1) … 2026-01-31.*
 
-- **Animartrix** overhauled — optional gamma correction, always paints in 2D, big math speedups, dependency upgrade, and several bugfixes (segment-option changes now respected).
+- **Animartrix** overhauled, optional gamma correction, always paints in 2D, big math speedups, dependency upgrade, and several bugfixes (segment-option changes now respected).
 - **New ESP32 node types** for ESP-NOW (WizMote data); Philips Hue robustness; PixelForge GIF tool gains image rotation.
 - **Fixed:** DMX-output now rate-limited to prevent watchdog resets; "relay does not turn on" sporadic issue; stack-smashing crash risk from `notify()`; better 2D preview color accuracy and PS Fireworks trails.
 - New V4 build environments incl. `esp32_16MB_V4_M_eth` (16 MB ESP32 with Ethernet); IR re-enabled for the Athom Music build.
 
 ## January 2026 (up to v14.7.1)
 
-*Summarised from 32 first-parent commits on `mdev`, 2026-01-01 … 2026-01-13 — released as **v14.7.1**.*
+*Summarised from 32 first-parent commits on `mdev`, 2026-01-01 … 2026-01-13, released as **v14.7.1**.*
 
 - **Release v14.7.1.**
 - Random per-LED colors via the JSON API (`"col":["r","r","r"]`); manual/dual auto-white modes work with palettes; segment-palette functions inlined for speed.
@@ -144,4 +144,4 @@ Near-dormant month — a single small change and a build-number bump.
 
 *Summarised from 3 first-parent commits on `mdev`, 2025-09-01 … 2025-09-30.*
 
-- Quiet month on `mdev` — build instructions, npm `ci` for dependencies, and GitHub Copilot contributor instructions. Nothing user-facing.
+- Quiet month on `mdev`, build instructions, npm `ci` for dependencies, and GitHub Copilot contributor instructions. Nothing user-facing.
diff --git a/docs/friend-repos/PlummersSoftwareLLC-NightDriverStrip.md b/docs/friend-repos/PlummersSoftwareLLC-NightDriverStrip.md
index 0abd95ab..fadf06dc 100644
--- a/docs/friend-repos/PlummersSoftwareLLC-NightDriverStrip.md
+++ b/docs/friend-repos/PlummersSoftwareLLC-NightDriverStrip.md
@@ -1,8 +1,8 @@
-# NightDriverStrip — monthly activity digest
+# NightDriverStrip: monthly activity digest
 
-What landed on [NightDriverStrip](https://github.com/PlummersSoftwareLLC/NightDriverStrip)'s `main` branch, month by month. External-context reference — a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [NightDriverStrip](https://github.com/PlummersSoftwareLLC/NightDriverStrip)'s `main` branch, month by month. External-context reference, a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
-Summarised via the GitHub commits API (no local clone), so counts are all commits on `main`, not first-parent merges — the bullets filter out dependency bumps, whitespace, and pure refactors. Releases are noted as context rather than used as month boundaries: **v1.3.0** (published 2026-01-10) was tagged from a late-November commit, and the latest are **v2.0.0** and **v2.0.1**, both published 2026-06-14. v1.3.0 is not a clean month boundary so its month is kept whole; June IS split at v2.0.0, which was cut from `main` mid-month (see the two June sections below).
+Summarised via the GitHub commits API (no local clone), so counts are all commits on `main`, not first-parent merges, the bullets filter out dependency bumps, whitespace, and pure refactors. Releases are noted as context rather than used as month boundaries: **v1.3.0** (published 2026-01-10) was tagged from a late-November commit, and the latest are **v2.0.0** and **v2.0.1**, both published 2026-06-14. v1.3.0 is not a clean month boundary so its month is kept whole; June IS split at v2.0.0, which was cut from `main` mid-month (see the two June sections below).
 
 ## August 2026
 
@@ -30,11 +30,11 @@ A quiet month: one feature merge, no release, no issues.
 - Better behaviour on low-memory boards: the firmware degrades gracefully instead of failing when memory runs short.
 - Fixed the JPEG decoder not being ready in time for the startup splash, and corrected serial status output on S3 boards.
 
-_Auditability: 3 commits on `main` author-dated 2026-07-01..2026-07-31 (1 first-parent merge — PR #901, merged July 18 — plus a whitespace commit); `commits?sha=main&since=…&until=…`. Issues checked via `search/issues` for `created:2026-07-01..2026-07-31` (0), `closed:` (0) and `updated:` (0) in the same range. No versioned release published in July (latest are v2.0.0/v2.0.1, both June 14), so the month is kept whole._
+_Auditability: 3 commits on `main` author-dated 2026-07-01..2026-07-31 (1 first-parent merge, PR #901, merged July 18, plus a whitespace commit); `commits?sha=main&since=…&until=…`. Issues checked via `search/issues` for `created:2026-07-01..2026-07-31` (0), `closed:` (0) and `updated:` (0) in the same range. No versioned release published in July (latest are v2.0.0/v2.0.1, both June 14), so the month is kept whole._
 
 ## June 2026 (up to v2.0.0)
 
-The big one: **NightDriverStrip 2.0.0** shipped on June 14 — a major release cut from `main`.
+The big one: **NightDriverStrip 2.0.0** shipped on June 14, a major release cut from `main`.
 
 **New**
 - Brand-new Web UI replacing the old one, plus a new browser-based web installer for flashing devices.
@@ -57,12 +57,12 @@ The big one: **NightDriverStrip 2.0.0** shipped on June 14 — a major release c
 
 - **v2.0.1** (also June 14): patch that fixes the web installer failing to build for the 2.0.0 release (removed a stale project entry).
 - Stock-ticker effect now shows correct live data, fetched through the new V2 API.
-- Fixed 64x32 (wide-and-short) displays that were rotating and doubling their content instead of scaling — effect previews in the Web UI and CLI now match the active output driver's pixel mapping (issue #878).
+- Fixed 64x32 (wide-and-short) displays that were rotating and doubling their content instead of scaling, effect previews in the Web UI and CLI now match the active output driver's pixel mapping (issue #878).
 
 **Watching**
 - Issue #877 ("networking seems broken") drew heavy discussion (38 comments) around the WebUI being unreachable after a recent merge on some setups; closed in June.
 
-Auditability: ~40 first-parent merges/commits on `main` with author-date in 2026-06-01..2026-06-30 (`commits?sha=main&since=…until=…`); two versioned releases published June 14 — v2.0.0 (commit ce00eaa) and v2.0.1 (commit 835015b), both ancestors of `main`, so the month is split at v2.0.0. Issues checked via `search/issues` for `created:2026-06-01..2026-06-30` (0 opened in range) and `closed:2026-06-01..2026-06-30` (#877, #878, #825 closed; #878 and #877 user-facing).
+Auditability: ~40 first-parent merges/commits on `main` with author-date in 2026-06-01..2026-06-30 (`commits?sha=main&since=…until=…`); two versioned releases published June 14, v2.0.0 (commit ce00eaa) and v2.0.1 (commit 835015b), both ancestors of `main`, so the month is split at v2.0.0. Issues checked via `search/issues` for `created:2026-06-01..2026-06-30` (0 opened in range) and `closed:2026-06-01..2026-06-30` (#877, #878, #825 closed; #878 and #877 user-facing).
 
 ## May 2026
 
@@ -70,13 +70,13 @@ Auditability: ~40 first-parent merges/commits on `main` with author-date in 2026
 
 **New**
 
-- **RGBWW / SK6812 white-channel support** — W/WW helpers, a WarmGlow test effect, configurable SK6812 white extraction; color-temperature naming cleanup.
-- **New Setup Wizard / guided-installer WebUI** — and a push to make the UI a "non-special consumer": everything the official UI needs now comes from the firmware over the wire (spec/schema), not baked into `app.js`.
+- **RGBWW / SK6812 white-channel support**: W/WW helpers, a WarmGlow test effect, configurable SK6812 white extraction; color-temperature naming cleanup.
+- **New Setup Wizard / guided-installer WebUI**: and a push to make the UI a "non-special consumer": everything the official UI needs now comes from the firmware over the wire (spec/schema), not baked into `app.js`.
 - M5 Stick S3 support (IR + WS2812B); optimised WS2812B draw path; `ACTIVITY_PIN` support; allow zero effects in the table.
 
 **Fixed / hardened**
 
-- Major **PSRAM strategy reversal** — switched to PSRAM-default routing (threshold 96, Mesmerizer's proven value) and removed the bespoke `psram_allocator` family; JSON save-path now uses internal RAM to avoid touching PSRAM during flash/cache-disabled windows.
+- Major **PSRAM strategy reversal**, switched to PSRAM-default routing (threshold 96, Mesmerizer's proven value) and removed the bespoke `psram_allocator` family; JSON save-path now uses internal RAM to avoid touching PSRAM during flash/cache-disabled windows.
 - Several concurrency and memory-safety fixes across buffer / drawing / network / task paths; PolarMap and noise-generation bugs; weather-data state protection.
 
 ## April 2026
@@ -96,7 +96,7 @@ Auditability: ~40 first-parent merges/commits on `main` with author-date in 2026
 
 *~32 commits on `main`, 2026-03-01 … 2026-03-31.*
 
-- **Replaced the RemoteDebug dependency** with a custom Logger + Telnet server — plus WiFi-stability and hardware-safety improvements; fixed a DebugCLI use-after-free for active telnet sessions and a stack-corruption (FD_SETSIZE) in the telnet sink.
+- **Replaced the RemoteDebug dependency** with a custom Logger + Telnet server, plus WiFi-stability and hardware-safety improvements; fixed a DebugCLI use-after-free for active telnet sessions and a stack-corruption (FD_SETSIZE) in the telnet sink.
 - New Arduino-V3 partition layouts (standard, NOOTA, 8MB).
 - Otherwise dependency bumps and include hygiene.
 
@@ -139,13 +139,13 @@ Auditability: ~40 first-parent merges/commits on `main` with author-date in 2026
 
 *2 commits on `main`, 2025-11-01 … 2025-11-30.*
 
-- Quiet month — a dependency bump (js-yaml). (The v1.3.0 tag was cut from a late-November commit but published in January.)
+- Quiet month, a dependency bump (js-yaml). (The v1.3.0 tag was cut from a late-November commit but published in January.)
 
 ## October 2025
 
 *2 commits on `main`, 2025-10-01 … 2025-10-31.*
 
-- Quiet month — a dependency bump (vite).
+- Quiet month, a dependency bump (vite).
 
 ## September 2025
 
diff --git a/docs/friend-repos/hpwit-ESPLiveScript.md b/docs/friend-repos/hpwit-ESPLiveScript.md
index be00870a..26a9d2aa 100644
--- a/docs/friend-repos/hpwit-ESPLiveScript.md
+++ b/docs/friend-repos/hpwit-ESPLiveScript.md
@@ -1,8 +1,8 @@
-# hpwit/ESPLiveScript — monthly activity digest
+# hpwit/ESPLiveScript: monthly activity digest
 
-What landed on [hpwit/ESPLiveScript](https://github.com/hpwit/ESPLiveScript), month by month. External-context reference — a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [hpwit/ESPLiveScript](https://github.com/hpwit/ESPLiveScript), month by month. External-context reference, a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
-The library: Yves Bazin's (hpwit) C-like compiler/interpreter for the ESP32 — small scripts (e.g. LED effects) compiled and run live on-device without a full recompile-and-flash cycle. Summarised via the GitHub commits API.
+The library: Yves Bazin's (hpwit) C-like compiler/interpreter for the ESP32, small scripts (e.g. LED effects) compiled and run live on-device without a full recompile-and-flash cycle. Summarised via the GitHub commits API.
 
 **Branch note:** `main` is quiet (last touched June 2025), but this repo develops on a long series of **version branches** (`v2`…`v4.3`, plus `vjson`/`vjson2`/`vdrop`/`memory*`), and that's where the recent work is. The activity below is read across those branches, not just `main`.
 
@@ -16,21 +16,21 @@ _Checked: `repos/hpwit/ESPLiveScript/commits?sha=main` for 2026-08-01..2026-09-0
 
 ## July 2026
 
-No user-facing activity: no commits on `main` **or any of the 38 version branches** (v2.x/v3.x/v4.x, `vjson`/`vjson2`/`vdrop`, `dev`, `mem*`) in July 2026, and no notable issues. (Latest commit on `main` predates the window — June 2025; the newest commit anywhere is `vjson2`, February 2026.)
+No user-facing activity: no commits on `main` **or any of the 38 version branches** (v2.x/v3.x/v4.x, `vjson`/`vjson2`/`vdrop`, `dev`, `mem*`) in July 2026, and no notable issues. (Latest commit on `main` predates the window, June 2025; the newest commit anywhere is `vjson2`, February 2026.)
 
-_Checked: commits author-dated 2026-07-01..2026-07-31 on `main` and every one of the 38 branches — 0 on each; issues created / closed / updated 2026-07-01..2026-07-31 (0 each); PRs created in-window (0); no versioned release published in July 2026._
+_Checked: commits author-dated 2026-07-01..2026-07-31 on `main` and every one of the 38 branches, 0 on each; issues created / closed / updated 2026-07-01..2026-07-31 (0 each); PRs created in-window (0); no versioned release published in July 2026._
 
 ## June 2026
 
-No user-facing activity: no commits on `main` **or any of the ~30 version branches** (v2.x/v3.x, dev, mem*) in June 2026, and no notable issues. (Latest commit on `main` predates the window — June 2025.)
+No user-facing activity: no commits on `main` **or any of the ~30 version branches** (v2.x/v3.x, dev, mem*) in June 2026, and no notable issues. (Latest commit on `main` predates the window, June 2025.)
 
-_Checked: commits author-dated 2026-06-01..2026-06-30 on `main` and every version branch (`v2`…`v3.3`, `dev`, `mem2`…`mem4`, `memory`) — 0 on each; issues created / closed / updated 2026-06-01..2026-06-30 (0 each); PRs created in-window (0); no versioned release published in June 2026._
+_Checked: commits author-dated 2026-06-01..2026-06-30 on `main` and every version branch (`v2`…`v3.3`, `dev`, `mem2`…`mem4`, `memory`), 0 on each; issues created / closed / updated 2026-06-01..2026-06-30 (0 each); PRs created in-window (0); no versioned release published in June 2026._
 
 ## February 2026
 
 *Latest in-window activity, on the `vjson2` branch.*
 
-- **JSON exchange refinements** (`vjson2`) — continued work on the script↔host JSON path begun in mid-2025, plus a code-refactoring cleanup pass.
+- **JSON exchange refinements** (`vjson2`), continued work on the script↔host JSON path begun in mid-2025, plus a code-refactoring cleanup pass.
 
 ## March 2025 (and earlier 2025, on the `v4.x` branches)
 
diff --git a/docs/friend-repos/hpwit-I2SClocklessLedDriver.md b/docs/friend-repos/hpwit-I2SClocklessLedDriver.md
index 7469d77e..970ebcd9 100644
--- a/docs/friend-repos/hpwit-I2SClocklessLedDriver.md
+++ b/docs/friend-repos/hpwit-I2SClocklessLedDriver.md
@@ -1,10 +1,10 @@
-# hpwit/I2SClocklessLedDriver — monthly activity digest
+# hpwit/I2SClocklessLedDriver: monthly activity digest
 
-What landed on [hpwit/I2SClocklessLedDriver](https://github.com/hpwit/I2SClocklessLedDriver)'s `main` branch, month by month. External-context reference — a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [hpwit/I2SClocklessLedDriver](https://github.com/hpwit/I2SClocklessLedDriver)'s `main` branch, month by month. External-context reference, a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
-The library: Yves Bazin's (hpwit) clockless-LED driver that clocks WS2812-class strips out over the ESP32's I2S/LCD peripheral with DMA — the parallel-output technique projectMM's LED-driver analysis studies. Summarised via the GitHub commits API (all commits on `main`), filtering out merge commits, lint/format churn, and 🐰-review fixups. No versioned release is cut from `main` in this window (latest tag is 1.4), so months are kept whole.
+The library: Yves Bazin's (hpwit) clockless-LED driver that clocks WS2812-class strips out over the ESP32's I2S/LCD peripheral with DMA, the parallel-output technique projectMM's LED-driver analysis studies. Summarised via the GitHub commits API (all commits on `main`), filtering out merge commits, lint/format churn, and 🐰-review fixups. No versioned release is cut from `main` in this window (latest tag is 1.4), so months are kept whole.
 
-> **Authorship note.** Most of the activity in this window is projectMM's own — `ewowi` authored ~53 of the in-window commits, with the rest from the maintainer (Yves Bazin / hpwit) and a couple of others. The IDF 5.5 / arduino-less ESP-IDF / RGBCCT / >65K-LED work below is largely projectMM upstreaming its driver needs into hpwit's library, then tracking the result here.
+> **Authorship note.** Most of the activity in this window is projectMM's own, `ewowi` authored ~53 of the in-window commits, with the rest from the maintainer (Yves Bazin / hpwit) and a couple of others. The IDF 5.5 / arduino-less ESP-IDF / RGBCCT / >65K-LED work below is largely projectMM upstreaming its driver needs into hpwit's library, then tracking the result here.
 
 ## August 2026
 
@@ -16,7 +16,7 @@ _Checked: `repos/hpwit/I2SClocklessLedDriver/commits?sha=main` for 2026-08-01..2
 
 ## July 2026
 
-No user-facing activity: no commits merged to `main` (latest activity is April 6, 2026) and no notable issues. No branch saw commits either — the newest work anywhere is the `esp32-p4-support` branch, last touched April 11, 2026.
+No user-facing activity: no commits merged to `main` (latest activity is April 6, 2026) and no notable issues. No branch saw commits either, the newest work anywhere is the `esp32-p4-support` branch, last touched April 11, 2026.
 
 _Auditability: commits on `main` author-dated 2026-07-01..2026-07-31 = 0 (0 merged), and 0 on every other branch; issues created/closed/updated in July 2026 = 0; PRs created = 0. No versioned release published in July (latest tag `1.4`, 2026-04-06)._
 
@@ -44,7 +44,7 @@ _Auditability: commits on `main` author-dated 2026-06-01..2026-06-30 = 0 (0 merg
 
 *~4 commits on `main`, 2026-01-01 … 2026-01-31.*
 
-- **>65K-LED support** — `total_leds` widened to 32 bits; added `extractWhiteFromRGB`.
+- **>65K-LED support**: `total_leds` widened to 32 bits; added `extractWhiteFromRGB`.
 - ESP32-D0: removed `ESP_INTR_FLAG_IRAM` from `esp_intr_alloc`; `deleteDriver` checks on the DMA tampon buffers.
 
 ## December 2025
@@ -65,7 +65,7 @@ _Auditability: commits on `main` author-dated 2026-06-01..2026-06-30 = 0 (0 merg
 
 **New**
 
-- **IDF 5.5 support** — version checks, dynamic DMA-buffer allocation (PSRAM-preferred), `NUM_STRIPS` as a global, `IRAM_ATTR` removed from forwards to compile warning-free.
+- **IDF 5.5 support**: version checks, dynamic DMA-buffer allocation (PSRAM-preferred), `NUM_STRIPS` as a global, `IRAM_ATTR` removed from forwards to compile warning-free.
 - `updateDriver` / `deleteDriver` gained length/size and per-strip offset parameters; `initled` with custom color arrangement; `isVirtualDriver` flag; `setDelay()`.
 - Split a `Driver.cpp` out of the header; added clang-format and removed the `COLOR_ORDER_` compiler directives.
 
diff --git a/docs/friend-repos/hpwit-I2SClocklessVirtualLedDriver.md b/docs/friend-repos/hpwit-I2SClocklessVirtualLedDriver.md
index 062b7331..89bc4b89 100644
--- a/docs/friend-repos/hpwit-I2SClocklessVirtualLedDriver.md
+++ b/docs/friend-repos/hpwit-I2SClocklessVirtualLedDriver.md
@@ -1,8 +1,8 @@
-# hpwit/I2SClocklessVirtualLedDriver — monthly activity digest
+# hpwit/I2SClocklessVirtualLedDriver: monthly activity digest
 
-What landed on [hpwit/I2SClocklessVirtualLedDriver](https://github.com/hpwit/I2SClocklessVirtualLedDriver), month by month. External-context reference — a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [hpwit/I2SClocklessVirtualLedDriver](https://github.com/hpwit/I2SClocklessVirtualLedDriver), month by month. External-context reference, a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
-The library: Yves Bazin's (hpwit) "virtual pins" variant of the I2S clockless driver — drives far more strips than the chip has usable pins by fanning the I2S output through external shift registers. This multiplex technique is the load-bearing idea projectMM's LED-driver analysis singles out (factoring the shift-register multiplex out of the I2S/LCD peripheral code). Summarised via the GitHub commits API, read across all branches (`main`, `integration`, `int2`, `variable`, `hpwit-patch-1`, `dev`, `optomize`), not just `main`.
+The library: Yves Bazin's (hpwit) "virtual pins" variant of the I2S clockless driver, drives far more strips than the chip has usable pins by fanning the I2S output through external shift registers. This multiplex technique is the load-bearing idea projectMM's LED-driver analysis singles out (factoring the shift-register multiplex out of the I2S/LCD peripheral code). Summarised via the GitHub commits API, read across all branches (`main`, `integration`, `int2`, `variable`, `hpwit-patch-1`, `dev`, `optomize`), not just `main`.
 
 ## August 2026
 
@@ -16,20 +16,20 @@ _Checked: `repos/hpwit/I2SClocklessVirtualLedDriver/commits?sha=main` for 2026-0
 
 No user-facing activity this month: no commits merged to `main` (latest commit on `main` dates to November 2024), no commits on any other branch (newest anywhere is `variable`, December 2024), no releases published, and no issues opened, closed, or updated.
 
-_Checked: commits with author-date 2026-07-01..2026-07-31 on `main` and every branch (`integration`, `int2`, `variable`, `hpwit-patch-1`, `dev`, `optomize`) — 0 on each; issues created 2026-07-01..2026-07-31 (0), closed in that range (0), and updated in that range (0); PRs created (0); releases (none in July — latest versioned tag is 2.1, Jan 2024)._
+_Checked: commits with author-date 2026-07-01..2026-07-31 on `main` and every branch (`integration`, `int2`, `variable`, `hpwit-patch-1`, `dev`, `optomize`), 0 on each; issues created 2026-07-01..2026-07-31 (0), closed in that range (0), and updated in that range (0); PRs created (0); releases (none in July, latest versioned tag is 2.1, Jan 2024)._
 
 ## June 2026
 
 No user-facing activity this month: no commits merged to `main` (latest commit on `main` dates to November 2024), no releases published, and no issues opened, closed, or updated.
 
-_Checked: commits on `main` with author-date 2026-06-01..2026-06-30 (0 commits); issues created 2026-06-01..2026-06-30 (0), closed in that range (0), and updated in that range (0); releases (none in June — latest versioned tag is 2.1, Jan 2024)._
+_Checked: commits on `main` with author-date 2026-06-01..2026-06-30 (0 commits); issues created 2026-06-01..2026-06-30 (0), closed in that range (0), and updated in that range (0); releases (none in June, latest versioned tag is 2.1, Jan 2024)._
 
 ## No activity in the digest window (Sept 2025 – May 2026)
 
 No branch has any commits in the window these digests cover; the project went quiet at the end of 2024. Its last active stretch:
 
-- **December 2024** (`variable` / `int2` branches) — variable-driver and integration work; this is the newest commit anywhere in the repo (`variable`, 2024-12-17).
-- **November 2024** (`main`, tag `2.1`) — **ESP32-S3 support** added and merged to `main` (2024-11-20).
-- **Earlier 2024** — `i2sStop` moved into IRAM (June); ESP-IDF 5 corrections (June); shift-array support and examples (Jan–Oct).
+- **December 2024** (`variable` / `int2` branches), variable-driver and integration work; this is the newest commit anywhere in the repo (`variable`, 2024-12-17).
+- **November 2024** (`main`, tag `2.1`), **ESP32-S3 support** added and merged to `main` (2024-11-20).
+- **Earlier 2024**: `i2sStop` moved into IRAM (June); ESP-IDF 5 corrections (June); shift-array support and examples (Jan–Oct).
 
-*Nothing newer to report. This file gets a new `## <Month Year>` section the next time any branch sees in-window commits — watch `integration`/`variable`, where the development historically happened before merging to `main`.*
+*Nothing newer to report. This file gets a new `## <Month Year>` section the next time any branch sees in-window commits, watch `integration`/`variable`, where the development historically happened before merging to `main`.*
diff --git a/docs/friend-repos/hpwit-new-parser.md b/docs/friend-repos/hpwit-new-parser.md
index 201d1a9b..3dc14778 100644
--- a/docs/friend-repos/hpwit-new-parser.md
+++ b/docs/friend-repos/hpwit-new-parser.md
@@ -1,10 +1,10 @@
-# hpwit/new-parser (ESPLiveScript2) — monthly activity digest
+# hpwit/new-parser (ESPLiveScript2): monthly activity digest
 
-What landed on [hpwit/new-parser](https://github.com/hpwit/new-parser), month by month. External-context reference — a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [hpwit/new-parser](https://github.com/hpwit/new-parser), month by month. External-context reference, a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
-The library: **ESPLiveScript2**, Yves Bazin's (hpwit) from-scratch C++ rewrite of [ESPLiveScript](https://github.com/hpwit/ESPLiveScript) — the same idea (a small C-like language compiled on-device to real Xtensa machine code, no interpreter, so a script runs at near-native speed) reimplemented independently rather than refactored. The library ships inside the repo as `asmparser2/` (PlatformIO name `ESPLiveScript2`, at v1.3.0). Summarised via the GitHub commits API.
+The library: **ESPLiveScript2**, Yves Bazin's (hpwit) from-scratch C++ rewrite of [ESPLiveScript](https://github.com/hpwit/ESPLiveScript), the same idea (a small C-like language compiled on-device to real Xtensa machine code, no interpreter, so a script runs at near-native speed) reimplemented independently rather than refactored. The library ships inside the repo as `asmparser2/` (PlatformIO name `ESPLiveScript2`, at v1.3.0). Summarised via the GitHub commits API.
 
-**Repo note:** the repository name is `new-parser`, but the library and its README call it **ESPLiveScript2** — the name to search for. Sibling digest for v1: [hpwit-ESPLiveScript.md](hpwit-ESPLiveScript.md).
+**Repo note:** the repository name is `new-parser`, but the library and its README call it **ESPLiveScript2**, the name to search for. Sibling digest for v1: [hpwit-ESPLiveScript.md](hpwit-ESPLiveScript.md).
 
 ## August 2026
 
@@ -64,7 +64,7 @@ _Checked: 38 commits author-dated 2026-08-01..2026-08-31 on `main` (0b6aa6b..d84
 
 ## Timeline note (added 2026-08-06)
 
-Added to the digest set on 2026-08-06, after the product owner flagged the rewrite. History to date, from the commit log: created March 2025, six commits across March–May 2025, then **dormant for over a year**, then **12 commits in the first days of August 2026** — the rewrite as it now stands is days old at the time of writing. July 2026 is therefore empty, and the August work is summarised above.
+Added to the digest set on 2026-08-06, after the product owner flagged the rewrite. History to date, from the commit log: created March 2025, six commits across March–May 2025, then **dormant for over a year**, then **12 commits in the first days of August 2026**, the rewrite as it now stands is days old at the time of writing. July 2026 is therefore empty, and the August work is summarised above.
 
 What the rewrite is, from its README (context for future months, not an endorsement):
 
@@ -74,6 +74,6 @@ What the rewrite is, from its README (context for future months, not an endorsem
 
 ## July 2026
 
-No activity: no commits on `main` in July 2026, and no issues. (The repo was dormant between May 2025 and August 2026 — the current rewrite work begins 2026-08-01, outside this window.)
+No activity: no commits on `main` in July 2026, and no issues. (The repo was dormant between May 2025 and August 2026, the current rewrite work begins 2026-08-01, outside this window.)
 
 _Checked: commits author-dated 2026-07-01..2026-07-31 on `main` (0); issues created 2026-07-01..2026-07-31 (0) and closed in the same window (0); no versioned release published in July 2026._
diff --git a/docs/friend-repos/index.md b/docs/friend-repos/index.md
index f225786f..37b2d963 100644
--- a/docs/friend-repos/index.md
+++ b/docs/friend-repos/index.md
@@ -1,40 +1,40 @@
 # Friend repos, monthly activity digests
 
-Monthly logs of what shipped on related open-source LED projects — the live landscape projectMM watches to sharpen its own designs under the *Industry standards, our own code* principle ([CLAUDE.md § Principles](../../CLAUDE.md#principles)): study to think, write fresh, never copy. Generated by the [digest prompt](#digest-prompt-reusable) below.
-
-- [FastLED-FastLED.md](FastLED-FastLED.md) — the LED-animation library; ESP32/Arduino driver + color math.
-- [wled-WLED.md](wled-WLED.md) — upstream WLED firmware.
-- [MoonModules-WLED-MM.md](MoonModules-WLED-MM.md) — MoonModules' WLED fork (the direct lineage).
-- [troyhacks-WLED.md](troyhacks-WLED.md) — troyhacks' personal fork of WLED-MM (PixelForge, RMTHI, audio-reactive hardening).
-- [Funkelfetisch-projectMM.md](Funkelfetisch-projectMM.md) — a fork of THIS project building a commercial product on it (HELIO, a physical infinity-sphere lamp); the work lives in feature branches, not on its default branch.
-- [PlummersSoftwareLLC-NightDriverStrip.md](PlummersSoftwareLLC-NightDriverStrip.md) — Dave Plummer's LED matrix/strip firmware.
-- [hpwit-I2SClocklessLedDriver.md](hpwit-I2SClocklessLedDriver.md) — hpwit's I2S/LCD DMA clockless LED driver (parallel multi-strip output).
-- [hpwit-I2SClocklessVirtualLedDriver.md](hpwit-I2SClocklessVirtualLedDriver.md) — the shift-register "virtual pins" variant of the above (dormant since 2024).
-- [hpwit-ESPLiveScript.md](hpwit-ESPLiveScript.md) — hpwit's live C-like script compiler for the ESP32 (main quiet; work moved to version branches).
-- [hpwit-new-parser.md](hpwit-new-parser.md) — **ESPLiveScript2**, hpwit's from-scratch rewrite of the above (repo is named `new-parser`; the library lives in `asmparser2/`). Dormant May 2025 → August 2026, then an active rewrite whose stated goal is a *verifiable* compiler: host builds plus QEMU running the actual compiled Xtensa bytes.
+Monthly logs of what shipped on related open-source LED projects, the live landscape projectMM watches to sharpen its own designs under the *Industry standards, our own code* principle ([CLAUDE.md § Principles](../../CLAUDE.md#principles)): study to think, write fresh, never copy. Generated by the [digest prompt](#digest-prompt-reusable) below.
+
+- [FastLED-FastLED.md](FastLED-FastLED.md): the LED-animation library; ESP32/Arduino driver + color math.
+- [wled-WLED.md](wled-WLED.md): upstream WLED firmware.
+- [MoonModules-WLED-MM.md](MoonModules-WLED-MM.md): MoonModules' WLED fork (the direct lineage).
+- [troyhacks-WLED.md](troyhacks-WLED.md): troyhacks' personal fork of WLED-MM (PixelForge, RMTHI, audio-reactive hardening).
+- [Funkelfetisch-projectMM.md](Funkelfetisch-projectMM.md): a fork of THIS project building a commercial product on it (HELIO, a physical infinity-sphere lamp); the work lives in feature branches, not on its default branch.
+- [PlummersSoftwareLLC-NightDriverStrip.md](PlummersSoftwareLLC-NightDriverStrip.md): Dave Plummer's LED matrix/strip firmware.
+- [hpwit-I2SClocklessLedDriver.md](hpwit-I2SClocklessLedDriver.md): hpwit's I2S/LCD DMA clockless LED driver (parallel multi-strip output).
+- [hpwit-I2SClocklessVirtualLedDriver.md](hpwit-I2SClocklessVirtualLedDriver.md): the shift-register "virtual pins" variant of the above (dormant since 2024).
+- [hpwit-ESPLiveScript.md](hpwit-ESPLiveScript.md): hpwit's live C-like script compiler for the ESP32 (main quiet; work moved to version branches).
+- [hpwit-new-parser.md](hpwit-new-parser.md): **ESPLiveScript2**, hpwit's from-scratch rewrite of the above (repo is named `new-parser`; the library lives in `asmparser2/`). Dormant May 2025 → August 2026, then an active rewrite whose stated goal is a *verifiable* compiler: host builds plus QEMU running the actual compiled Xtensa bytes.
 
 ## Cross-repo trends
 
 Reading across the friend-repo digests, the themes the wider ESP32-LED ecosystem converged on over this release cycle (Sept 2025 → June 2026):
 
 - **ESP32-P4 / S3 parallel output.** FastLED poured effort into the PARLIO and LCD_CAM drivers (P4/S3 parallel LED output, big encode speedups); NightDriverStrip added custom RMT output; hpwit's I2SClocklessLedDriver pushed IDF-5.5 + arduino-less-ESP-IDF support for its I2S/LCD DMA driver (the canonical implementation of this technique), and troyhacks ran ESP32-P4 bring-up branches. The frontier is parallel, DMA-driven output on the newer chips.
-- **PSRAM strategy is unsettled everywhere.** All four wrestled with PSRAM this cycle — WLED-MM moved preview buffers into PSRAM, NightDriverStrip did a full PSRAM-default reversal (then tuned the threshold), WLED added S3-no-PSRAM builds. Nobody has a clean answer; the cache-disabled-during-flash hazard recurs across repos.
+- **PSRAM strategy is unsettled everywhere.** All four wrestled with PSRAM this cycle, WLED-MM moved preview buffers into PSRAM, NightDriverStrip did a full PSRAM-default reversal (then tuned the threshold), WLED added S3-no-PSRAM builds. Nobody has a clean answer; the cache-disabled-during-flash hazard recurs across repos.
 - **Audio-reactive maturing.** FastLED added a silence-gate + ESP-DSP FFT backend; WLED-MM and WLED both refined audio sync and auto-disable-during-realtime; NightDriverStrip modernised its SoundAnalyzer/FFT. Audio-reactive is table stakes now, and the polish is in *not* reacting to noise/silence.
 - **The FastLED dependency question.** WLED merged a *full FastLED replacement* (its own color/math); projectMM already made the same call (own color math, no FastLED in core). Two independent projects concluded the dependency wasn't worth it.
-- **UI as a firmware-driven consumer.** Both WLED and NightDriverStrip pushed toward "the official UI knows nothing the firmware doesn't publish over the wire" — exactly projectMM's MoonModule-driven, no-hardcoded-knowledge UI principle. Convergent design. NightDriverStrip's **2.0.0** (June 2026) crystallised this: a brand-new web UI, a browser-based installer, and settings (like strip type) moved from compile-time to *runtime-selectable* on the device — the same "reconfigure live, no reflash" direction projectMM builds around.
+- **UI as a firmware-driven consumer.** Both WLED and NightDriverStrip pushed toward "the official UI knows nothing the firmware doesn't publish over the wire", exactly projectMM's MoonModule-driven, no-hardcoded-knowledge UI principle. Convergent design. NightDriverStrip's **2.0.0** (June 2026) crystallised this: a brand-new web UI, a browser-based installer, and settings (like strip type) moved from compile-time to *runtime-selectable* on the device, the same "reconfigure live, no reflash" direction projectMM builds around.
 - **Effect velocity.** WLED and WLED-MM shipped many new effects (PacMan, Color Clouds, Shimmer, the user_fx pack); new effects remain the most visible user-facing output.
-- **Display / HDMI output beyond LED strips.** troyhacks ran a cluster of branches probing HDMI video output and large hardware panels (WaveShare 10.1″, M5Stack, ESP32-P4 panels) — driving *displays*, not just addressable strips, off the same firmware.
-- **On-device live scripting.** hpwit's ESPLiveScript compiles small C-like effect scripts that run live on the ESP32 with no reflash — a different answer to effect authoring than C++ recompilation or a fixed effect table.
+- **Display / HDMI output beyond LED strips.** troyhacks ran a cluster of branches probing HDMI video output and large hardware panels (WaveShare 10.1″, M5Stack, ESP32-P4 panels), driving *displays*, not just addressable strips, off the same firmware.
+- **On-device live scripting.** hpwit's ESPLiveScript compiles small C-like effect scripts that run live on the ESP32 with no reflash, a different answer to effect authoring than C++ recompilation or a fixed effect table.
 
 ## What these projects do that projectMM doesn't (yet)
 
 Observational: where the landscape is ahead of projectMM. These are *not* commitments; real adoption decisions live in the [`../backlog/`](../work/future/index.md), cross-referenced where one already exists.
 
-- **Parallel multi-strip output on S3/P4** (PARLIO/LCD_CAM, and hpwit's I2S/shift-register drivers) — the direct parallel drivers ship (MultiPin/Moon on LCD_CAM, Parlio on P4, driving up to 16 strands and 12,288+ lights). The shift-register/'595 expander path also ships but is dormant: it works at prime-only geometries yet has a known lapping-ring sparkle at the largest configs, so it stays off by default. See the [LED-driver analysis](../work/future/leddriver-analysis-top-down.md).
-- **Audio-reactive input** — none of projectMM's effects are audio- or motion-reactive yet. The Peripheral role + the Pi-sensor backlog entry are the foundation; the producer→effect wiring is backlog.
-- **A guided setup/installer wizard on-device** (NightDriverStrip's Setup Wizard, WLED's installer) — projectMM has the web installer + Improv, but no on-device first-run wizard.
-- **A large built-in effect library** — projectMM ships a focused set (concrete-first); the WLED family ships dozens. Breadth is a deliberate non-goal until the core is proven.
-- **On-device live effect scripting** (hpwit's ESPLiveScript) — projectMM effects are compiled C++; there's no runtime script path. Not a goal today, noted as a landscape contrast.
+- **Parallel multi-strip output on S3/P4** (PARLIO/LCD_CAM, and hpwit's I2S/shift-register drivers), the direct parallel drivers ship (MultiPin/Moon on LCD_CAM, Parlio on P4, driving up to 16 strands and 12,288+ lights). The shift-register/'595 expander path also ships but is dormant: it works at prime-only geometries yet has a known lapping-ring sparkle at the largest configs, so it stays off by default. See the [LED-driver analysis](../work/future/leddriver-analysis-top-down.md).
+- **Audio-reactive input**: none of projectMM's effects are audio- or motion-reactive yet. The Peripheral role + the Pi-sensor backlog entry are the foundation; the producer→effect wiring is backlog.
+- **A guided setup/installer wizard on-device** (NightDriverStrip's Setup Wizard, WLED's installer), projectMM has the web installer + Improv, but no on-device first-run wizard.
+- **A large built-in effect library**: projectMM ships a focused set (concrete-first); the WLED family ships dozens. Breadth is a deliberate non-goal until the core is proven.
+- **On-device live effect scripting** (hpwit's ESPLiveScript), projectMM effects are compiled C++; there's no runtime script path. Not a goal today, noted as a landscape contrast.
 
 ## Refreshing
 
@@ -44,12 +44,13 @@ Adding a month or a new friend repo is the [friend-repos](index.md) workflow, an
 
 > **Friend-repo monthly digest.** For the repo `<NAME>` (local clone at `<PATH>`, or via `gh api repos/<owner>/<NAME>`), summarise what landed on its **main/default branch** during `<MONTH YEAR>`.
 >
-> 1. Read the merged commits on the default branch with author-date in that calendar month (`git log --first-parent --since/--until` on the local clone, or the GitHub API). Use `--first-parent` so it's the merged-feature view, not every squashed sub-commit. The default branch isn't always `main`/`master` — check (`git remote show origin`); e.g. WLED-MM's is `mdev`.
-> 1b. **Also investigate the issues over that month.** The REST `gh api repos/<owner>/<NAME>/issues` endpoint returns **pull requests too** — filter them out (`--jq '.[] | select(.pull_request == null)'`) or use the GitHub **search** API, which already excludes them: `search/issues?q=repo:<owner>/<NAME>+is:issue+created:<YYYY-MM-DD..YYYY-MM-DD>` (and the same with `closed:`). Only real issues, not PRs. The commit log shows what shipped; the issues show what users *hit* and what the maintainers are prioritising — the two together are the real activity picture. Skim: notable bugs opened (recurring pain points, hardware quirks), fixes closed that map to a commit, and any heavily-discussed feature request or design thread. Fold the user-facing ones into the summary below (a widely-reported bug that got fixed, a feature the community is pushing for); an issue with no user-facing outcome yet is still worth a one-line "watching:" note if it signals a direction. Don't list every issue — surface the few that matter, the same bar as the commit summary.
-> 2. **Split a month at any release boundary — but only if the release was cut from the branch you're summarising.** If a *versioned* release was published mid-month (check `git for-each-ref refs/tags` / the GitHub releases API; ignore rolling tags like `nightly` and prereleases), AND the tag is an ancestor of the digest branch (`git merge-base --is-ancestor <tag> <branch>`), split that month at the release date into `## <Month Year> (up to v<X>)` / `## <Month Year> (post-v<X>)`. If the tag is NOT an ancestor (the project cuts releases from a separate release branch — e.g. upstream WLED tags off `0_15`/release branches, not `main`), do NOT split: keep the month whole and just note which release shipped that month as context, since the trunk you're summarising feeds future releases rather than being the release line. Whole months with no in-branch release stay one section.
-> 3. Write an **end-user-readable** summary: what changed that a *user of the library* would notice or care about — new features, new hardware/platform support, notable fixes, breaking changes. Skip internal refactors, CI, test-only, and dependency bumps unless they affect users.
+> 1. Read the merged commits on the default branch with author-date in that calendar month (`git log --first-parent --since/--until` on the local clone, or the GitHub API). Use `--first-parent` so it's the merged-feature view, not every squashed sub-commit. The default branch isn't always `main`/`master`, check (`git remote show origin`); e.g. WLED-MM's is `mdev`.
+> 1b. **Also investigate the issues over that month.** The REST `gh api repos/<owner>/<NAME>/issues` endpoint returns **pull requests too**, filter them out (`--jq '.[] | select(.pull_request == null)'`) or use the GitHub **search** API, which already excludes them: `search/issues?q=repo:<owner>/<NAME>+is:issue+created:<YYYY-MM-DD..YYYY-MM-DD>` (and the same with `closed:`). Only real issues, not PRs. The commit log shows what shipped; the issues show what users *hit* and what the maintainers are prioritising, the two together are the real activity picture. Skim: notable bugs opened (recurring pain points, hardware quirks), fixes closed that map to a commit, and any heavily-discussed feature request or design thread. Fold the user-facing ones into the summary below (a widely-reported bug that got fixed, a feature the community is pushing for); an issue with no user-facing outcome yet is still worth a one-line "watching:" note if it signals a direction. Don't list every issue, surface the few that matter, the same bar as the commit summary.
+> 2. **Split a month at any release boundary, but only if the release was cut from the branch you're summarising.** If a *versioned* release was published mid-month (check `git for-each-ref refs/tags` / the GitHub releases API; ignore rolling tags like `nightly` and prereleases), AND the tag is an ancestor of the digest branch (`git merge-base --is-ancestor <tag> <branch>`), split that month at the release date into `## <Month Year> (up to v<X>)` / `## <Month Year> (post-v<X>)`. If the tag is NOT an ancestor (the project cuts releases from a separate release branch, e.g. upstream WLED tags off `0_15`/release branches, not `main`), do NOT split: keep the month whole and just note which release shipped that month as context, since the trunk you're summarising feeds future releases rather than being the release line. Whole months with no in-branch release stay one section.
+> 3. Write an **end-user-readable** summary: what changed that a *user of the library* would notice or care about, new features, new hardware/platform support, notable fixes, breaking changes. Skip internal refactors, CI, test-only, and dependency bumps unless they affect users.
 > 4. Format as **short bullet points**, each one line, plainest language, minimal jargon. Group only if there's a natural split (e.g. "New" / "Fixed"); otherwise a flat list.
-> 5. Add it as a `## <MONTH YEAR>` section to `docs/friend-repos/<NAME>.md`, newest month on top. Don't editorialise or compare to projectMM — just report what they shipped.
+> 4b. Follow [the documentation standards](../contributing/documentation-standards.md): American spelling, **no em-dashes** (use a comma, colon, parentheses or a full stop), and one line per paragraph with no hard wraps, since `check_docgen` reports all three as errors on the file you write.
+> 5. Add it as a `## <MONTH YEAR>` section to `docs/friend-repos/<NAME>.md`, newest month on top. Don't editorialise or compare to projectMM, just report what they shipped.
 > 6. State the commit range / count **and the issue query** summarised so the digest is auditable.
 >
-> When backfilling several months (e.g. since the last release), run this once per month for a consistent timeline, then optionally add a `## Since v<last-release> — overview` intro at the top with 3–5 bullets naming the multi-month threads the per-month slices can't show on their own.
+> When backfilling several months (e.g. since the last release), run this once per month for a consistent timeline, then optionally add a `## Since v<last-release>, overview` intro at the top with 3–5 bullets naming the multi-month threads the per-month slices can't show on their own.
diff --git a/docs/friend-repos/troyhacks-WLED.md b/docs/friend-repos/troyhacks-WLED.md
index b1fbcb4b..67930c43 100644
--- a/docs/friend-repos/troyhacks-WLED.md
+++ b/docs/friend-repos/troyhacks-WLED.md
@@ -1,10 +1,10 @@
-# troyhacks/WLED — monthly activity digest
+# troyhacks/WLED: monthly activity digest
 
-What landed on [troyhacks/WLED](https://github.com/troyhacks/WLED)'s `mdev` branch, month by month. External-context reference — a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [troyhacks/WLED](https://github.com/troyhacks/WLED)'s `mdev` branch, month by month. External-context reference, a factual log of a friend repo's activity, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
 This is a personal fork of [MoonModules/WLED-MM](https://github.com/MoonModules/WLED-MM) (the `mdev` branch), so much of the `mdev` traffic is merges from and alignments with the MM and upstream WLED lines; the bullets below pick out what a *user* of this fork would notice. Summarised via the GitHub commits API (all commits on `mdev`, not first-parent merges), filtering out build-number bumps, merge commits, typo/comment churn, and pure refactors. No versioned release is cut from this branch (it tracks `mdev` and ships nightly `mdev` builds), so months are kept whole.
 
-**Branch note — the experiments live off `mdev`.** troyhacks branches heavily: `mdev` is the merge/alignment stream, but the distinctive work happens in named experimental branches (HDMI output, ESP32-P4, W5500 Ethernet, hardware-panel ports, voice control, a pure-IDFv5 port, a new settings subsystem). Those are *experiments*, not necessarily destined for `mdev`, so each month below carries a separate **Experimental branches** line for what moved on them — the frontier of what this fork is probing.
+**Branch note, the experiments live off `mdev`.** troyhacks branches heavily: `mdev` is the merge/alignment stream, but the distinctive work happens in named experimental branches (HDMI output, ESP32-P4, W5500 Ethernet, hardware-panel ports, voice control, a pure-IDFv5 port, a new settings subsystem). Those are *experiments*, not necessarily destined for `mdev`, so each month below carries a separate **Experimental branches** line for what moved on them, the frontier of what this fork is probing.
 
 ## August 2026
 
@@ -18,22 +18,22 @@ _Checked: commits on `mdev` for author-date 2026-08-01..2026-08-31 (2: b537e0c9
 
 No user-facing activity: no commits were merged to `mdev` in July 2026 (the branch's most recent commit is still dated 2026-05-20), and no versioned release was published. The repository's issue tracker is disabled, so no issues were opened or closed.
 
-- **Experimental branches:** nothing moved in July either — the most-recently-touched branch, `P4_experimental` (ESP32-P4), had no July commit either, and neither did any other branch.
+- **Experimental branches:** nothing moved in July either, the most-recently-touched branch, `P4_experimental` (ESP32-P4), had no July commit either, and neither did any other branch.
 
-_Checked: merged commits on `mdev` for author-date 2026-07-01..2026-08-01 (0 commits); commits on `P4_experimental` for the same window (0); releases published in July 2026 (none); issue search `repo:troyhacks/WLED is:issue created:2026-07-01..2026-07-31` and `closed:2026-07-01..2026-07-31` (0 results — issues disabled on this fork)._
+_Checked: merged commits on `mdev` for author-date 2026-07-01..2026-08-01 (0 commits); commits on `P4_experimental` for the same window (0); releases published in July 2026 (none); issue search `repo:troyhacks/WLED is:issue created:2026-07-01..2026-07-31` and `closed:2026-07-01..2026-07-31` (0 results, issues disabled on this fork)._
 
 ## June 2026
 
 No user-facing activity: no commits were merged to `mdev` in June 2026 (the branch's most recent commit is dated 2026-05-20), and no versioned release was published. The repository's issue tracker is disabled, so no issues were opened or closed.
 
-_Checked: merged commits on `mdev` for author-date 2026-06-01..2026-06-30 (0 commits); releases published in June 2026 (none); issue search `repo:troyhacks/WLED is:issue created:2026-06-01..2026-06-30` and `closed:2026-06-01..2026-06-30` (0 results — issues disabled on this fork)._
+_Checked: merged commits on `mdev` for author-date 2026-06-01..2026-06-30 (0 commits); releases published in June 2026 (none); issue search `repo:troyhacks/WLED is:issue created:2026-06-01..2026-06-30` and `closed:2026-06-01..2026-06-30` (0 results, issues disabled on this fork)._
 
 ## May 2026
 
 *~18 commits on `mdev`, 2026-05-01 … 2026-05-31.*
 
-- Quieter month on `mdev` — alignment with upstream/MM, smaller effect and build-flag fixes; the heavy lifting was in the Nov–Jan window.
-- **Experimental branches:** `P4_experimental` (ESP32-P4) and `M5Stack_Core_S3_Display` both saw work — the two most-recently-touched branches in the repo.
+- Quieter month on `mdev`, alignment with upstream/MM, smaller effect and build-flag fixes; the heavy lifting was in the Nov–Jan window.
+- **Experimental branches:** `P4_experimental` (ESP32-P4) and `M5Stack_Core_S3_Display` both saw work, the two most-recently-touched branches in the repo.
 
 ## April 2026
 
@@ -49,7 +49,7 @@ _Checked: merged commits on `mdev` for author-date 2026-06-01..2026-06-30 (0 com
 
 - Effect and 2D-matrix fixes; build-target and partition adjustments.
 - More upstream-compatibility alignment in the segment / bus drawing code.
-- **Experimental branches:** `Pure_IDFv5_Port` (a from-scratch ESP-IDF v5 port — no Arduino) and `New-Settings-Subsystem` (settings rework with usermod auto-detection) both started.
+- **Experimental branches:** `Pure_IDFv5_Port` (a from-scratch ESP-IDF v5 port, no Arduino) and `New-Settings-Subsystem` (settings rework with usermod auto-detection) both started.
 
 ## February 2026
 
@@ -86,14 +86,14 @@ _Checked: merged commits on `mdev` for author-date 2026-06-01..2026-06-30 (0 com
 
 **New**
 
-- **WLEDPixelForge** — a new image and scrolling-text interface (`pxmagic`), with 1D GIF support, blur option, and version-14.x adaptations.
+- **WLEDPixelForge**: a new image and scrolling-text interface (`pxmagic`), with 1D GIF support, blur option, and version-14.x adaptations.
 - Effect math sped up (up to ~3× faster); inlined hot-path color/segment functions; more segment/effect data allowed on PSRAM boards.
 - DDP-over-websockets / DDP-over-WS stability; E1.31 kill switch; `dnrgbw` realtime mode.
 
 **Fixed / hardened**
 
 - Large **preset/ledmap robustness pass**: fixed `presets.json` corruption (mutex protection on the write path), ledmap-parser robustness, reduced UI freeze when updating presets.
-- Extensive **mutex / critical-section redesign** across segment and `bus.show` paths — fixes for semaphore leaks, "giving a semaphore never taken", and realtime-lock race conditions.
+- Extensive **mutex / critical-section redesign** across segment and `bus.show` paths, fixes for semaphore leaks, "giving a semaphore never taken", and realtime-lock race conditions.
 
 ## November 2025
 
diff --git a/docs/friend-repos/wled-WLED.md b/docs/friend-repos/wled-WLED.md
index f24a586f..9bb43201 100644
--- a/docs/friend-repos/wled-WLED.md
+++ b/docs/friend-repos/wled-WLED.md
@@ -1,8 +1,8 @@
-# WLED (upstream) — monthly activity digest
+# WLED (upstream): monthly activity digest
 
-What landed on [wled/WLED](https://github.com/wled/WLED)'s `main` branch, month by month. External-context reference — a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
+What landed on [wled/WLED](https://github.com/wled/WLED)'s `main` branch, month by month. External-context reference, a factual log of a friend repo's releases, not projectMM's own history or roadmap. Newest month on top. The reusable prompt that generates these lives in [README.md](index.md).
 
-Months are **not** split at release dates: upstream WLED cuts releases from separate release branches (`0_15`, `16_x`), so the version tags aren't on `main` — `main` is the development trunk that feeds future releases. Each month notes which release shipped, as context.
+Months are **not** split at release dates: upstream WLED cuts releases from separate release branches (`0_15`, `16_x`), so the version tags aren't on `main`, `main` is the development trunk that feeds future releases. Each month notes which release shipped, as context.
 
 ## August 2026
 
@@ -41,7 +41,7 @@ The month `main` switched to the **V5** platform: WLED's trunk moved from the ES
 
 **New**
 - Trunk builds move to ESP-IDF 5.3 / arduino-esp32 v3, opening the door to the newer chips (ESP32-C5, C6 and P4 build targets ride along).
-- ESP-NOW now uses WLED's own code instead of the QuickESPNow library — faster, and roughly 10 KB more free memory on ESP32 (1.5 KB on ESP8266).
+- ESP-NOW now uses WLED's own code instead of the QuickESPNow library, faster, and roughly 10 KB more free memory on ESP32 (1.5 KB on ESP8266).
 - New `esp32_eth_V4` build for Ethernet boards; ESP32-C6 boards get 4 MB and 8 MB builds.
 - Nightly builds renamed, and the web UI warns when you're running a development build.
 - Audio-reactive now compiles on all the newer chips.
@@ -81,7 +81,7 @@ Post-16.0 stabilisation month: no new version tag (v16.0.0 shipped 2026-05-03 of
 - Better brownout detection and extended error codes aligned with WLED-MM.
 
 **Watching**
-- Discussion opened on switching from plain gamma to an sRGB transfer function for better low-brightness accuracy (#5707), and on improving the Nodes/Instances page (#5711) — no shipped outcome yet.
+- Discussion opened on switching from plain gamma to an sRGB transfer function for better low-brightness accuracy (#5707), and on improving the Nodes/Instances page (#5711), no shipped outcome yet.
 - Several v16.0 field reports still open: multi-controller sync losing color (#5705), UDP sync failing in AP mode (#5709), and OTA-update trouble on some boards (#5682, #5702).
 
 _Auditability: 43 commits on `main` with author-date 2026-06-01..2026-06-30 (`repos/wled/WLED/commits?sha=main`, first-line view; a few older-dated cherry-picks appear in-range and were excluded as non-June). Issues via `search/issues` for repo:wled/WLED created:2026-06-01..2026-06-30 (18 opened) and closed:2026-06-01..2026-06-30 (25 closed); only user-facing ones surfaced. No versioned release published in June (v16.0.0 was 2026-05-03), so no month split._
@@ -121,7 +121,7 @@ _Auditability: 43 commits on `main` with author-date 2026-06-01..2026-06-30 (`re
 
 **New / effects**
 
-- **Full FastLED replacement** merged (#4615) — WLED's own color/math instead of the FastLED dependency.
+- **Full FastLED replacement** merged (#4615), WLED's own color/math instead of the FastLED dependency.
 - Many new user_fx effects: Spinning Wheel, Color Clouds, Lava Lamp, Magma, Ants, Morse Code, Comet (fire particle system), a slow >4-hour transition FX, Tetris line-clear flash.
 - Scrolling-text FX gains custom fonts + international UTF-8; stencil blending mode; ESP32-C3 audio-reactive (DSP FFT + integer math); more macro/timer slots; longer max playlist duration.
 - OTA update page restyled (auto-sets download URL from `info.repo`); clearer UI tool icons.
@@ -138,7 +138,7 @@ _Auditability: 43 commits on `main` with author-date 2026-06-01..2026-06-30 (`re
 
 - **Version scheme changed to Major.minor** (dropped the leading "0."), heading toward v16; bumped to 16.0.0-alpha.
 - New **Pin Info** page (used/available pins overview); UI settings readability improvements.
-- Improved bus handling — free choice of bus driver in any order, better memory calculations; gamma lower-limit removed (enables inverse gamma correction, applied to segment brightness too).
+- Improved bus handling, free choice of bus driver in any order, better memory calculations; gamma lower-limit removed (enables inverse gamma correction, applied to segment brightness too).
 - Extended CCT blending (exclusive blend, color-jump fix); full WiFi scan with BSSID apply; new ESP32-S3 8MB QSPI build; experimental ESP32-C5/C6 in the node list.
 
 **Fixed**
diff --git a/docs/gettingstarted.md b/docs/gettingstarted.md
index 324ae774..764dfffd 100644
--- a/docs/gettingstarted.md
+++ b/docs/gettingstarted.md
@@ -1,18 +1,13 @@
 # Getting started
 
-New to ESP32 or flashing firmware? You don't need to be. projectMM installs
-straight from your web browser: no software to download, no command line. In a
-few minutes you'll have lights running and the device on your network, and the
-device's own web interface open in your browser ready to play with.
+New to ESP32 or flashing firmware? You don't need to be. projectMM installs straight from your web browser: no software to download, no command line. In a few minutes you'll have lights running and the device on your network, and the device's own web interface open in your browser ready to play with.
 
 This guide has two chapters. **Chapter 1** gets projectMM onto your device.
 **Chapter 2** is a tour of the interface you land in afterwards, so you know what
 every part does and where to start building your own light show.
 
 **You need:** an ESP32 board, a USB cable that carries data (not charge-only),
-and a **Chromium-based browser** on a computer (Google Chrome, Microsoft Edge,
-or Opera (the installer uses the Web Serial API, which Safari and Firefox don't
-support).
+and a **Chromium-based browser** on a computer (Google Chrome, Microsoft Edge, or Opera (the installer uses the Web Serial API, which Safari and Firefox don't support).
 
 > Want the bigger picture of what projectMM is first? See the
 > [project overview](../README.md).
@@ -23,8 +18,7 @@ support).
 
 ### 1. Open the installer and plug in
 
-Open the **[web installer](https://moonmodules.org/projectMM/install/)** in
-Chrome or Edge, then plug your ESP32 into a USB port.
+Open the **[web installer](https://moonmodules.org/projectMM/install/)** in Chrome or Edge, then plug your ESP32 into a USB port.
 
 ![The web installer](assets/gettingstarted/01-01-installer-start.png)
 
@@ -32,70 +26,49 @@ Chrome or Edge, then plug your ESP32 into a USB port.
 
 ### 2. Pick the USB port
 
-Click **USB Port → Pick a port…**. Your browser shows a small list of connected
-devices — choose the one that appeared when you plugged in the ESP32. (Not sure
-which? Unplug, look at the list, plug back in — the new entry is your device.)
+Click **USB Port → Pick a port…**. Your browser shows a small list of connected devices, choose the one that appeared when you plugged in the ESP32. (Not sure which? Unplug, look at the list, plug back in, the new entry is your device.)
 
 ![Selecting the USB port](assets/gettingstarted/01-02-select-port.png)
 
-**Windows users — dialog says "No serial ports available"?** Windows doesn't
-ship drivers for the USB-serial chips most ESP32 boards use (WCH CH340, Silicon
-Labs CP2102). One-time install fixes it for every future flash — full
-step-by-step + the download link is in
-[building.md § Windows: USB-serial drivers](how-to/building.md#windows-usb-serial-drivers).
+**Windows users, dialog says "No serial ports available"?** Windows doesn't
+ship drivers for the USB-serial chips most ESP32 boards use (WCH CH340, Silicon Labs CP2102). One-time install fixes it for every future flash, full step-by-step + the download link is in [building.md § Windows: USB-serial drivers](how-to/building.md#windows-usb-serial-drivers).
 macOS and Linux ship these drivers built in, so it's a Windows-only step.
 
-Once a port is chosen, the installer recognises the chip and tells you how many
-devices match it, so you know you're on the right track before you pick one.
+Once a port is chosen, the installer recognizes the chip and tells you how many devices match it, so you know you're on the right track before you pick one.
 
 ![Port selected, chip detected](assets/gettingstarted/01-03-port-selected.png)
 
 ### 3. Pick your device
 
-Choose your device from the **Device** picker. Each card shows a picture, the
-chip, and what the device can do (LEDs, WiFi, a button, a microphone…); click
+Choose your device from the **Device** picker. Each card shows a picture, the chip, and what the device can do (LEDs, WiFi, a button, a microphone…); click
 **details** on any card to see exactly what it is and a link to its product page.
 
 ![Picking a device](assets/gettingstarted/01-04-pick-device.png)
 
 ![A device card with its details](assets/gettingstarted/01-05-device-details.png)
 
-The little colored pills are the device's capabilities, and the color tells you
-how ready each one is:
+The little colored pills are the device's capabilities, and the color tells you how ready each one is:
 
-- 🟢 **Green** — set up and working the moment you install. This capability is
+- 🟢 **Green**, set up and working the moment you install. This capability is
   supported *and* already wired into the device's configuration.
-- 🟡 **Yellow** — the firmware supports it, but it isn't pre-configured. It works
+- 🟡 **Yellow**, the firmware supports it, but it isn't pre-configured. It works
   once you add and set up the matching module yourself in the UI (Chapter 2).
-- 🟠 **Amber** — planned. The hardware has it, but there's no module for it yet —
-  it's on the to-do list. (Want to help? Building one is our usual loop: read the
-  product page and datasheet, pin the behavior as tests, then write the code to
-  pass them — [see how we work](../CLAUDE.md#principles).)
+- 🟠 **Amber**, planned. The hardware has it, but there's no module for it yet: it's on the to-do list. (Want to help? Building one is our usual loop: read the product page and datasheet, pin the behavior as tests, then write the code to pass them, [see how we work](../CLAUDE.md#principles).)
 
-So a green pill is "just works", a yellow one is "works, with a bit of setup", and
-an amber one is "coming later".
+So a green pill is "just works", a yellow one is "works, with a bit of setup", and an amber one is "coming later".
 
-The setup panel then shows how your device is configured out of the box — the
-modules and settings applied automatically when you install.
+The setup panel then shows how your device is configured out of the box, the modules and settings applied automatically when you install.
 
 ![Device setup](assets/gettingstarted/01-06-device-setup.png)
 
-Nothing is locked in: once the device is running you can change any of it later
-in the UI (that's what Chapter 2 is all about).
+Nothing is locked in: once the device is running you can change any of it later in the UI (that's what Chapter 2 is all about).
 
-Leave **Release** and **Firmware** at their suggested values (the newest stable
-build, and the firmware that matches your device). Tick **Erase chip first** only
-if you're starting clean, switching firmware, or updating a 4 MB classic board
-(esp32 / wrover / eth) from a release before v4.0. That last update must erase:
-its partition layout changed ([MIGRATING](reference/MIGRATING.md)), and if the device already holds
-config you care about, back it up first ("Back up a device's config first" on the
-installer page): erasing wipes WiFi credentials and all settings, and the backup
-brings them back after the flash (its report lists anything it could not carry).
+Leave **Release** and **Firmware** at their suggested values (the newest stable build, and the firmware that matches your device). Tick **Erase chip first** only if you're starting clean, switching firmware, or updating a 4 MB classic board (esp32 / wrover / eth) from a release before v4.0. That last update must erase:
+its partition layout changed ([MIGRATING](reference/MIGRATING.md)), and if the device already holds config you care about, back it up first ("Back up a device's config first" on the installer page): erasing wipes WiFi credentials and all settings, and the backup brings them back after the flash (its report lists anything it could not carry).
 
 ### 4. Click Install
 
-The installer erases (if you asked it to) and writes the firmware. Just watch —
-it takes under a minute.
+The installer erases (if you asked it to) and writes the firmware. Just watch, it takes under a minute.
 
 ![Erasing](assets/gettingstarted/01-07-erasing.png)
 ![Installing](assets/gettingstarted/01-08-installing.png)
@@ -105,40 +78,30 @@ it takes under a minute.
 What happens next depends on your device:
 
 - **WiFi:** enter your network name and password when prompted, then **Connect**.
-  (Click **Skip** to set WiFi up later from the device itself.) Restoring a config
-  backup? You can skip this step: join the device's `MM-XXXX` access point, open
-  `http://4.3.2.1`, and restore the backup in the File Manager (⟲), then take the
-  offered restart: the bundle carries the WiFi credentials, so the device joins
-  your network by itself.
+  (Click **Skip** to set WiFi up later from the device itself.) Restoring a config backup? You can skip this step: join the device's `MM-XXXX` access point, open `http://4.3.2.1`, and restore the backup in the File Manager (⟲), then take the offered restart: the bundle carries the WiFi credentials, so the device joins your network by itself.
 
   ![Entering WiFi credentials](assets/gettingstarted/01-09-wifi-credentials.png)
 
-- **Ethernet:** plug in the cable — it connects on its own, no password needed.
+- **Ethernet:** plug in the cable, it connects on its own, no password needed.
 
 ### 6. Open your device
 
-When it's online, the installer shows a link — your device's address on your
-network. Click it.
+When it's online, the installer shows a link, your device's address on your network. Click it.
 
 ![Device is online over WiFi](assets/gettingstarted/01-10-online-wifi.png)
 
-You'll see this same "Device is online!" box however your device connected — over
-Ethernet, or when it rejoins a network it already knows:
+You'll see this same "Device is online!" box however your device connected, over Ethernet, or when it rejoins a network it already knows:
 
 ![Online over Ethernet](assets/gettingstarted/01-11-online-ethernet.png)
 ![Online on an address it already had](assets/gettingstarted/01-12-online-existing-ip.png)
 
-That's it — projectMM is installed and on your network. The link opens the
-device's own web interface, served straight from the ESP32. Let's look around.
+That's it, projectMM is installed and on your network. The link opens the device's own web interface, served straight from the ESP32. Let's look around.
 
 ---
 
-## Chapter 2 — Your projectMM interface
+## Chapter 2: Your projectMM interface
 
-Everything below runs **in your browser, live from the device**. There's no app,
-no account, no cloud — the ESP32 itself serves this page, and every change you
-make takes effect on the lights immediately. Open the link from step 6 and follow
-along; you can't break anything by exploring.
+Everything below runs **in your browser, live from the device**. There's no app, no account, no cloud, the ESP32 itself serves this page, and every change you make takes effect on the lights immediately. Open the link from step 6 and follow along; you can't break anything by exploring.
 
 ### The layout: list, preview, controls
 
@@ -146,86 +109,68 @@ along; you can't break anything by exploring.
 
 Three regions, left to right:
 
-- **The module list** (left) — every part of your device, from system info at the
+- **The module list** (left), every part of your device, from system info at the
   top to your light setup at the bottom. Click a name to jump to it.
 - **The 3D preview** (center): a live picture of your lights in their real shape,
   updating as the effects run. This is what your physical LEDs are doing, right now.
-- **The controls** (right) — the settings for each module. Drag a slider or pick an
+- **The controls** (right), the settings for each module. Drag a slider or pick an
   option and the lights react instantly.
 
-Every module header carries a **⏻ power button** — it turns that module on or off.
-Bright (accent-colored) means on; dimmed means off. A switched-off module simply
-stops running — it stays in place with all its settings, so flicking it back on
-picks up right where it left off. It's the quick way to mute an effect or an output
-for a moment without deleting anything.
+Every module header carries a **⏻ power button**, it turns that module on or off.
+Bright (accent-colored) means on; dimmed means off. A switched-off module simply stops running, it stays in place with all its settings, so flicking it back on picks up right where it left off. It's the quick way to mute an effect or an output for a moment without deleting anything.
 
-You'll also spot two little read-outs in each header: **🕒** is how fast that module
-runs (its loop speed — click it to flip between fps and microseconds), and **🧠** is
-how much memory it uses. They let you see at a glance what each part is costing.
+You'll also spot two little read-outs in each header: **🕒** is how fast that module runs (its loop speed, click it to flip between fps and microseconds), and **🧠** is how much memory it uses. They let you see at a glance what each part is costing.
 
-The interface adapts to your window. On a narrower screen the controls take the
-full width and the preview tucks into a floating thumbnail you can move around:
+The interface adapts to your window. On a narrower screen the controls take the full width and the preview tucks into a floating thumbnail you can move around:
 
-![Medium width — preview as a floating thumbnail](assets/gettingstarted/02-02-UI-mid.png)
+![Medium width, preview as a floating thumbnail](assets/gettingstarted/02-02-UI-mid.png)
 
-Narrower still, it stacks into a single scrollable column — so it works on a
-phone, standing next to your lights:
+Narrower still, it stacks into a single scrollable column, so it works on a phone, standing next to your lights:
 
-![Small width — single column](assets/gettingstarted/02-03-UI-small.png)
+![Small width, single column](assets/gettingstarted/02-03-UI-small.png)
 
 ### The 3D preview
 
 ![The 3D preview, lights numbered](assets/gettingstarted/02-04-UI-Preview.png)
 
-Drag to rotate, scroll to zoom. Each dot is one light at its real position, lit
-with the color it's showing this instant. Turn on the numbers to see each light's
-index — handy when you're wiring or mapping a layout. The preview is a *view* of
-the device; it never slows the lights down, and it gracefully eases off (fewer
-updates, then fewer points) on a slow connection rather than stalling.
+Drag to rotate, scroll to zoom. Each dot is one light at its real position, lit with the color it's showing this instant. Turn on the numbers to see each light's index, handy when you're wiring or mapping a layout. The preview is a *view* of the device; it never slows the lights down, and it gracefully eases off (fewer updates, then fewer points) on a slow connection rather than stalling.
 
 > More on how the preview streams from the device:
 > [PreviewDriver](moonmodules/light/moxygen/PreviewDriver.md).
 
 ### The system modules
 
-The top of the list is your device's "about" section — read-outs and connection
-settings. You rarely need to touch these, but they're the first place to look if
-something seems off.
+The top of the list is your device's "about" section, read-outs and connection settings. You rarely need to touch these, but they're the first place to look if something seems off.
 
-**System** — who this device is and how it's doing: its name, the device model,
-uptime, frame rate, and live memory / storage bars. You may also see an **Audio**
-module here — devices with a built-in mic come with it set up for you, and on any
-device you can add it yourself (it's how audio-reactive effects hear the music).
-Audio is just the first of many: any sensor or input — from hardware or over the
-network — lives here as its own module, and we're adding more all the time.
+**System**, who this device is and how it's doing: its name, the device model,
+uptime, frame rate, and live memory / storage bars. You may also see an **Audio** module here, devices with a built-in mic come with it set up for you, and on any device you can add it yourself (it's how audio-reactive effects hear the music).
+Audio is just the first of many: any sensor or input, from hardware or over the network, lives here as its own module, and we're adding more all the time.
 
 ![The System module](assets/gettingstarted/02-05-UI-System.png)
 
 > [SystemModule](moonmodules/core/system.md#system) ·
 > [Audio](moonmodules/core/services.md#audio)
 
-**Firmware** — which build you're running, and where you update it. The
-**Install** button here does an over-the-air update straight from the device — no
+**Firmware**, which build you're running, and where you update it. The
+**Install** button here does an over-the-air update straight from the device, no
 USB cable needed once it's on your network.
 
 ![The Firmware module](assets/gettingstarted/02-06-UI-Firmware.png)
 
-**Updating from an older build?** Skim the [migration notes](reference/MIGRATING.md) first. Most updates need nothing — the device keeps your settings — but a breaking change is listed there with the one action it costs you (usually re-setting or re-adding a control).
+**Updating from an older build?** Skim the [migration notes](reference/MIGRATING.md) first. Most updates need nothing, the device keeps your settings, but a breaking change is listed there with the one action it costs you (usually re-setting or re-adding a control).
 
 > [FirmwareUpdateModule](moonmodules/core/system.md#firmware-update)
 
-**Network** — your connection: WiFi or Ethernet, signal strength, and the
-address others reach it at. The **Devices** section underneath finds other
-projectMM devices on the same network, so a roomful of them can discover each
-other.
+**Network**, your connection: WiFi or Ethernet, signal strength, and the
+address others reach it at. The **Devices** section underneath finds other projectMM devices on the same network, so a roomful of them can discover each other.
 
 ![The Network module](assets/gettingstarted/02-07-UI-Network.png)
 
 > [NetworkModule](moonmodules/core/system.md#network) ·
 > [DevicesModule](moonmodules/core/system.md#devices)
 
-> **Lights are just one use.** Everything above — the modules, the live controls, the
-> 3D view, the web UI, the networking — is a general-purpose engine that knows nothing
+> **Lights are just one use.** Everything above, the modules, the live controls, the
+> 3D view, the web UI, the networking, is a general-purpose engine that knows nothing
 > about LEDs. The light show below is one *domain* built on top of it; you could build
 > a different one and reuse all the same machinery. [FastLED-MM](https://github.com/MoonModules/FastLED-MM)
 > is an example, driving its LEDs with [FastLED](https://github.com/FastLED/FastLED) (on
@@ -233,12 +178,7 @@ other.
 
 ### Control it from your phone with WLED Native
 
-The device's own web UI works on a phone, but for quick on/off and brightness from
-your pocket there's a nicer option: **WLED Native**, the open-source mobile app for
-the WLED ecosystem. projectMM speaks the WLED JSON API and announces itself over the
-network the same way a WLED device does, so the app finds your projectMM controllers
-automatically — no setup, no pairing. Each one shows up as a card with a power toggle
-and a brightness slider, so a roomful of controllers is a scroll and a tap away.
+The device's own web UI works on a phone, but for quick on/off and brightness from your pocket there's a nicer option: **WLED Native**, the open-source mobile app for the WLED ecosystem. projectMM speaks the WLED JSON API and announces itself over the network the same way a WLED device does, so the app finds your projectMM controllers automatically, no setup, no pairing. Each one shows up as a card with a power toggle and a brightness slider, so a roomful of controllers is a scroll and a tap away.
 
 ![projectMM devices discovered in WLED Native](assets/core/WLED%20Native%20discovers%20projectMM.jpeg){ width="300" }
 
@@ -249,110 +189,81 @@ Get it free for your phone:
 
 WLED Native is by **Christophe Gagnier ([@Moustachauve](https://github.com/Moustachauve))**, who wrote both the [Android](https://github.com/Moustachauve/WLED-Android) and [iOS](https://github.com/Moustachauve/WLED-iOS) apps. Their open source is what let us work out exactly what those apps read, so a projectMM device appears in them without either side needing to know about the other.
 
-For the full picture and controls, the device's web interface is always there at
-`http://<devicename>.local` — WLED Native is the fast everyday remote alongside it.
+For the full picture and controls, the device's web interface is always there at `http://<devicename>.local`, WLED Native is the fast everyday remote alongside it.
 
 ### Bring it into your smart home with Home Assistant
 
-Want your lights in the same dashboard as the rest of your house — and in
-automations, voice assistants, and Apple Home? projectMM adopts into **Home
-Assistant** like any other light: point the device at your HA setup and it appears
-as a light entity with on/off and brightness, alongside a floor of other devices.
+Want your lights in the same dashboard as the rest of your house, and in automations, voice assistants, and Apple Home? projectMM adopts into **Home Assistant** like any other light: point the device at your HA setup and it appears as a light entity with on/off and brightness, alongside a floor of other devices.
 
 ![projectMM devices as lights in a Home Assistant dashboard](assets/core/ha-integration.png){ width="600" }
 
-There are two ways in — zeroconf (HA finds the device on its own) or MQTT
-auto-discovery (for a broker-only or cross-subnet setup) — and from there you can
-bridge the entity into Apple Home too. The step-by-step, including installing HA
-and the MQTT broker if you don't have them, is in the
-[home automation guide](how-to/home-automation.md).
+There are two ways in, zeroconf (HA finds the device on its own) or MQTT auto-discovery (for a broker-only or cross-subnet setup), and from there you can bridge the entity into Apple Home too. The step-by-step, including installing HA and the MQTT broker if you don't have them, is in the [home automation guide](how-to/home-automation.md).
 
 ### Building a light show: layouts → layers → drivers
 
 The bottom three modules are where the fun is. They form a simple pipeline: a
 **layout** says where your lights are, **layers** decide what colors play on
-them, and **drivers** send the result out to the real world. Add modules with the
-dashed **+ add module** button under each one.
+them, and **drivers** send the result out to the real world. Add modules with the dashed **+ add module** button under each one.
 
-**Layouts** — the shape of your lights. The default **Grid** is a width × height
-(× depth) of pixels; change the numbers and the preview reshapes instantly. Turn
-on **serpentine** if your strip zig-zags back and forth.
+**Layouts**, the shape of your lights. The default **Grid** is a width × height
+(× depth) of pixels; change the numbers and the preview reshapes instantly. Turn on **serpentine** if your strip zig-zags back and forth.
 
 ![The Layouts module](assets/gettingstarted/02-08-UI-Layouts.png)
 
 > [Layouts](moonmodules/light/supporting.md)
 
-**Effects** — what plays on the lights. Add an **effect** (a moving pattern), stack
-several to blend them, and reshape them with **modifiers** (mirror, rotate, and
-more). Each effect has its own controls — speed, color mode, and so on — that you
-tweak live.
+**Effects**, what plays on the lights. Add an **effect** (a moving pattern), stack
+several to blend them, and reshape them with **modifiers** (mirror, rotate, and more). Each effect has its own controls, speed, color mode, and so on, that you tweak live.
 
 ![The Effects module](assets/gettingstarted/02-09-UI-Layers.png)
 
 > [Effects](moonmodules/light/supporting.md) · [Layer](moonmodules/light/supporting.md)
 
-**Drivers** — where the colors go. Set overall **brightness** and color order,
-then add an output: real LED strips on a pin, or send the frame over the network
-(ArtNet, E1.31/sACN, DDP) to other devices or lighting software.
+**Drivers**, where the colors go. Set overall **brightness** and color order,
+then add an output: real LED strips on a pin, or send the frame over the network (ArtNet, E1.31/sACN, DDP) to other devices or lighting software.
 
 ![The Drivers module](assets/gettingstarted/02-10-UI-Drivers.png)
 
 > [Drivers](moonmodules/light/supporting.md) ·
 > [NetworkSendDriver](moonmodules/light/moxygen/NetworkSendDriver.md)
 
-That's the whole picture: **layout → layers → drivers**, previewed in 3D, all
-tuned live in your browser. Pick an effect, drag a slider, watch the lights — then
-keep going.
+That's the whole picture: **layout → layers → drivers**, previewed in 3D, all tuned live in your browser. Pick an effect, drag a slider, watch the lights, then keep going.
 
 ---
 
 ### If your device shows MoonBase
 
 **MoonBase** is a small recovery image built into your device. If a firmware update
-is interrupted, or an installed firmware does not start, your device boots MoonBase
-instead of going dark, and its page offers you three ways out:
+is interrupted, or an installed firmware does not start, your device boots MoonBase instead of going dark, and its page offers you three ways out:
 
 - **Boot the app** puts you straight back if the firmware is still fine. Try this
   first: it changes nothing on the device.
 - **From a file** installs a firmware you have already downloaded. Get the
-  `firmware-...bin` matching your device from the
-  [releases page](https://github.com/MoonModules/projectMM/releases).
+  `firmware-...bin` matching your device from the [releases page](https://github.com/MoonModules/projectMM/releases).
 - **From a URL** downloads and installs in one step. The releases page gives you a
   link to each file; paste it here and your device fetches it directly.
 
 Installing takes a few minutes, and the page reports its progress as it downloads.
 Your device reboots into the new firmware on its own when it finishes.
 
-Two things worth knowing. A failed install **stays** in MoonBase rather than
-pretending to have worked, so you can simply try again. And you cannot break a
-device this way: MoonBase is never overwritten by an update, so it is still there
-for the next attempt, including after a power cut in the middle of one.
+Two things worth knowing. A failed install **stays** in MoonBase rather than pretending to have worked, so you can simply try again. And you cannot break a device this way: MoonBase is never overwritten by an update, so it is still there for the next attempt, including after a power cut in the middle of one.
 
-If your device is not on your network at all, MoonBase opens its own WiFi access
-point and is reachable at **4.3.2.1** once you join it.
+If your device is not on your network at all, MoonBase opens its own WiFi access point and is reachable at **4.3.2.1** once you join it.
 
-MoonBase shows its own version on its page, and your device's Firmware card shows
-which MoonBase it carries. If that version is marked outdated, the same card installs
-a newer one over the network, so keeping the recovery image current needs no cable.
+MoonBase shows its own version on its page, and your device's Firmware card shows which MoonBase it carries. If that version is marked outdated, the same card installs a newer one over the network, so keeping the recovery image current needs no cable.
 
-That update runs from the app, because only the running app can write the partition
-MoonBase lives in. So it is a way to keep MoonBase fresh, not a way back from a
-device that will not start: if the app cannot run, or MoonBase itself will not boot,
-that still takes a cable. Your device checks the image first, refusing anything whose
-magic bytes, chip or description say it is not a MoonBase image for this chip.
+That update runs from the app, because only the running app can write the partition MoonBase lives in. So it is a way to keep MoonBase fresh, not a way back from a device that will not start: if the app cannot run, or MoonBase itself will not boot, that still takes a cable. Your device checks the image first, refusing anything whose magic bytes, chip or description say it is not a MoonBase image for this chip.
 
 ---
 
 ### Where to go next
 
-- **Understand the pipeline** — how layouts, layers, effects, modifiers and
+- **Understand the pipeline**: how layouts, layers, effects, modifiers and
   drivers fit together: [architecture overview](explanation/architecture/moonlight.md#the-pipeline).
-- **Run it on your computer** instead of (or alongside) an ESP32 — macOS, Windows,
+- **Run it on your computer** instead of (or alongside) an ESP32, macOS, Windows,
   Linux: [project overview → Getting started](../README.md#getting-started).
 - **Manage several devices, build, and flash from one console** with MoonDeck, our
   developer tool: [MoonDeck guide](../moondeck/MoonDeck.md).
 - **Build from source** or target Teensy / Raspberry Pi: [building.md](how-to/building.md).
 
-Stuck, or something didn't work? Open an
-[issue](https://github.com/MoonModules/projectMM/issues) — and tell us what device
-you used and where it stopped.
+Stuck, or something didn't work? Open an [issue](https://github.com/MoonModules/projectMM/issues), and tell us what device you used and where it stopped.
diff --git a/docs/how-to/building.md b/docs/how-to/building.md
index 6b3a6bbc..e131ffb7 100644
--- a/docs/how-to/building.md
+++ b/docs/how-to/building.md
@@ -55,25 +55,16 @@ The distinction matters because a shipped binary is launched from a download fol
 
 ### Editor setup (clangd)
 
-Diagnostics appear **as you type**, from the same [`.clang-tidy`](../.clang-tidy) config CI
-uses, so a finding shows up while the code is still in your head, not ten minutes later in a
-pipeline.
+Diagnostics appear **as you type**, from the same [`.clang-tidy`](../.clang-tidy) config CI uses, so a finding shows up while the code is still in your head, not ten minutes later in a pipeline.
 
-Once per machine: install the **clangd** extension (`llvm-vs-code-extensions.vscode-clangd`)
-and **disable Microsoft's C/C++ IntelliSense**, running both produces duplicated and
-contradictory diagnostics. Nothing else to configure: [`.clangd`](../.clangd) at the repo root
-points at the compilation database, and `CMAKE_EXPORT_COMPILE_COMMANDS` (set in
-`CMakeLists.txt`) means any normal build refreshes it.
+Once per machine: install the **clangd** extension (`llvm-vs-code-extensions.vscode-clangd`) and **disable Microsoft's C/C++ IntelliSense**, running both produces duplicated and contradictory diagnostics. Nothing else to configure: [`.clangd`](../.clangd) at the repo root points at the compilation database, and `CMAKE_EXPORT_COMPILE_COMMANDS` (set in `CMakeLists.txt`) means any normal build refreshes it.
 
 Two things worth knowing:
 
 - **If every file reports `'cstdint' file not found`**: the build directory was configured
-  with a different compiler than clangd is. `.clangd`'s `--query-driver` handles the usual
-  cases; if a new toolchain appears, add it there. This failure is loud and total, real
-  diagnostics disappear behind it, so it is worth recognizing on sight.
+  with a different compiler than clangd is. `.clangd`'s `--query-driver` handles the usual cases; if a new toolchain appears, add it there. This failure is loud and total, real diagnostics disappear behind it, so it is worth recognizing on sight.
 - **clangd runs a subset of the CI check set**: skipping checks it considers slow (>10%
-  AST-build cost). That is deliberate and means the same config file is safe to share: CI
-  remains the authority.
+  AST-build cost). That is deliberate and means the same config file is safe to share: CI remains the authority.
 
 ### Packaging
 
@@ -100,8 +91,7 @@ Every host needs [uv](https://docs.astral.sh/uv/), CMake 3.20+, and a C++20 comp
 
 ### Docker
 
-The desktop build runs in a container, which is the whole system without an ESP32: same effect
-pipeline, same web UI, same driver stack, driving real fixtures over Art-Net, DDP and E1.31.
+The desktop build runs in a container, which is the whole system without an ESP32: same effect pipeline, same web UI, same driver stack, driving real fixtures over Art-Net, DDP and E1.31.
 
 ```sh
 docker compose up -d      # then open http://localhost:8081/
@@ -116,20 +106,12 @@ docker run -d --name projectmm -p 8081:8080 -v projectmm:/data \
   ghcr.io/moonmodules/projectmm:latest --no-browser
 ```
 
-`--no-browser` because a container has no browser to open: without it the start prints a line
-saying it could not open one, which is noise rather than a failure. The Compose service passes it
-already.
+`--no-browser` because a container has no browser to open: without it the start prints a line saying it could not open one, which is noise rather than a failure. The Compose service passes it already.
 
-`:latest` follows the rolling prerelease, the same build the installer page offers; a version tag
-like `:4.0.0` pins one. Images are published by the release workflow from the same `.deb` that
-release ships, so the image and the binary are the same build.
+`:latest` follows the rolling prerelease, the same build the installer page offers; a version tag like `:4.0.0` pins one. Images are published by the release workflow from the same `.deb` that release ships, so the image and the binary are the same build.
 
 **Upgrading preserves everything.** The image holds only the binary and all state lives in the
-volume, so an upgrade keeps settings, presets, scripts and the device's identity. Which command
-depends on which source the compose file names: `docker compose pull && docker compose up -d` for
-a published `image:`, and `docker compose build --pull && docker compose up -d` for the shipped
-`build: .`, where `pull` alone would fetch nothing and `up` would rebuild from the checkout. Only `docker compose down -v` wipes it, and only a mounted volume is preserved
-at all: a bare `docker run` with no `-v` loses its state when the container goes.
+volume, so an upgrade keeps settings, presets, scripts and the device's identity. Which command depends on which source the compose file names: `docker compose pull && docker compose up -d` for a published `image:`, and `docker compose build --pull && docker compose up -d` for the shipped `build: .`, where `pull` alone would fetch nothing and `up` would rebuild from the checkout. Only `docker compose down -v` wipes it, and only a mounted volume is preserved at all: a bare `docker run` with no `-v` loses its state when the container goes.
 
 | | |
 |---|---|
@@ -141,19 +123,13 @@ at all: a bare `docker run` with no `-v` loses its state when the container goes
 | **Capabilities** | none; it binds its port as an ordinary process |
 
 **When host networking is needed.** Unicast output to a fixture works over ordinary bridge
-networking. Discovery and the broadcast or multicast output modes do not cross a bridge, so those
-want `network_mode: host` (or an L2 CNI on Kubernetes). With host networking there is no port
-mapping, so pass `--port 8081` in `command:` to stay clear of anything already on 8080.
+networking. Discovery and the broadcast or multicast output modes do not cross a bridge, so those want `network_mode: host` (or an L2 CNI on Kubernetes). With host networking there is no port mapping, so pass `--port 8081` in `command:` to stay clear of anything already on 8080.
 
 **Several instances** run side by side with no conflict: each container has its own port space, so
-they all listen on 8080 internally with different published ports, and each generates its own
-identity so they are distinguishable on the network. CPU is the practical limit rather than memory
-(measured at ~5 MB and about one core each, since the desktop build renders as fast as it is
-allowed); cap it with `cpus:` in compose when running a fleet.
+they all listen on 8080 internally with different published ports, and each generates its own identity so they are distinguishable on the network. CPU is the practical limit rather than memory (measured at ~5 MB and about one core each, since the desktop build renders as fast as it is allowed); cap it with `cpus:` in compose when running a fleet.
 
 **amd64 only** for now: the release ships no arm64 Linux binary. On an Apple-silicon Mac or an ARM
-server the compose file's `platform: linux/amd64` runs it under emulation, which works but is
-slower than native.
+server the compose file's `platform: linux/amd64` runs it under emulation, which works but is slower than native.
 
 ## ESP32
 
@@ -189,13 +165,9 @@ uv run moondeck/build/flash_esp32.py --firmware esp32 --port /dev/tty.usbserial-
 uv run moondeck/run/monitor_esp32.py --port /dev/tty.usbserial-XXXX
 ```
 
-On the variants that opt into it (`esp32`, `esp32-16mb`, `esp32-wrover`, `esp32-eth`,
-`esp32s3-zero`, and `qemu`, which is emulated rather than installable) the build also produces
+On the variants that opt into it (`esp32`, `esp32-16mb`, `esp32-wrover`, `esp32-eth`, `esp32s3-zero`, and `qemu`, which is emulated rather than installable) the build also produces
 **MoonBase**, the second boot image ([MoonBase](../explanation/architecture/moonbase.md)),
-and `flash_esp32.py` writes the corrected layout in one pass: app in the big `ota_0` slot,
-MoonBase in `factory`, and an otadata that boots the app directly. A device on the older
-dual-OTA table adopts this layout only through such a full serial flash, OTA never rewrites
-the partition table.
+and `flash_esp32.py` writes the corrected layout in one pass: app in the big `ota_0` slot, MoonBase in `factory`, and an otadata that boots the app directly. A device on the older dual-OTA table adopts this layout only through such a full serial flash, OTA never rewrites the partition table.
 
 `setup_esp_idf.py` runs the upstream installer for the host: `install.sh` on macOS/Linux, `install.bat` on Windows. Both create the same `~/.espressif/python_env/...` venv and download the same toolchains (~1.5 GB more) — only the wrapper differs. The Windows installer needs roughly 5 minutes on a fast link. It also offers to move a drifted checkout onto the pinned commit (see [ESP-IDF version](#esp-idf-version)); pass `--no-checkout` to keep it warn-only.
 
@@ -320,10 +292,7 @@ The Ethernet PHY type and pin map are runtime config, not baked into the build:
 
 ### Flashing a running device over the network
 
-A board already on the network is updated over HTTP, with no cable. Which route to use depends on
-whether the variant carries [MoonBase](../explanation/architecture/moonbase.md): a board
-cannot rewrite the partition it is executing from, so on a MoonBase variant the app hands over to
-MoonBase and MoonBase does the writing.
+A board already on the network is updated over HTTP, with no cable. Which route to use depends on whether the variant carries [MoonBase](../explanation/architecture/moonbase.md): a board cannot rewrite the partition it is executing from, so on a MoonBase variant the app hands over to MoonBase and MoonBase does the writing.
 
 **On a MoonBase variant** (`esp32`, `esp32-16mb`, `esp32s3-zero`, and any variant `build_esp32.py`
 builds MoonBase alongside), it is two requests:
@@ -337,37 +306,26 @@ curl --http1.1 -H "Expect:" --data-binary @build/esp32-<firmware>/projectMM.bin
      http://<device>/api/firmware/upload
 ```
 
-The second request ends with **no HTTP status** (curl reports 000): the device reboots into the new
-image as the write completes, so the socket closes before a response arrives. That is success, not
-failure. Confirm by reading the build back:
+The second request ends with **no HTTP status** (curl reports 000): the device reboots into the new image as the write completes, so the socket closes before a response arrives. That is success, not failure. Confirm by reading the build back:
 
 ```sh
 curl -s http://<device>/api/modules/Firmware   # the `build` control names the commit and date
 ```
 
-MoonBase serves the same route names as the application, so a page driving an update keeps calling
-the same paths after the handover. It also installs unattended from a URL, which is what the UI's
-update button uses: `POST /api/firmware/url` with the URL as the body. `POST /api/firmware/boot-app`
-returns to the application without installing anything, and only boots an image that validates.
+MoonBase serves the same route names as the application, so a page driving an update keeps calling the same paths after the handover. It also installs unattended from a URL, which is what the UI's update button uses: `POST /api/firmware/url` with the URL as the body. `POST /api/firmware/boot-app` returns to the application without installing anything, and only boots an image that validates.
 
 **Without MoonBase**, the application takes the image directly on the same route,
-`POST /api/firmware/upload`. It is one of only two streaming routes (`/api/file` is the other), so
-the body may exceed the request buffer; every other route rejects an oversized body with 413.
+`POST /api/firmware/upload`. It is one of only two streaming routes (`/api/file` is the other), so the body may exceed the request buffer; every other route rejects an oversized body with 413.
 
 Two failure modes are worth recognizing, because both look like something else:
 
 - **413 from `/api/firmware/upload`** on a MoonBase variant means the request reached the
-  APPLICATION rather than MoonBase, and the app rejected an oversized body on a route it does not
-  stream. The device did not reboot into MoonBase, or booted back before the upload. Check with
-  `GET /moonbase`, which MoonBase answers and the app 404s.
+  APPLICATION rather than MoonBase, and the app rejected an oversized body on a route it does not stream. The device did not reboot into MoonBase, or booted back before the upload. Check with `GET /moonbase`, which MoonBase answers and the app 404s.
 - **`{"error":"incomplete request body"}`** from `/api/firmware/upload` means the body did not
-  arrive within the read window. Send with `--http1.1 -H "Expect:"` so the transfer starts
-  immediately instead of waiting for a `100 Continue` the device does not send.
+  arrive within the read window. Send with `--http1.1 -H "Expect:"` so the transfer starts immediately instead of waiting for a `100 Continue` the device does not send.
 
 **A partition-table change needs a cable.** OTA writes the app, never the table, so a device on an
-older layout adopts a new one only through a full serial flash (see the note under
-[Firmware variants](#firmware-variants)). On the 4 MB classic that migration also moves the
-filesystem, so the device comes back unprovisioned.
+older layout adopts a new one only through a full serial flash (see the note under [Firmware variants](#firmware-variants)). On the 4 MB classic that migration also moves the filesystem, so the device comes back unprovisioned.
 
 
 ## Teensy
diff --git a/docs/how-to/logging-an-issue.md b/docs/how-to/logging-an-issue.md
index 42a4e256..8a47a125 100644
--- a/docs/how-to/logging-an-issue.md
+++ b/docs/how-to/logging-an-issue.md
@@ -1,9 +1,6 @@
 # Log an issue
 
-Something not working? Tell us. You do not need to diagnose it — describe what
-you saw, and hand us the state of the device it happened on. The **`{ }`** link
-on every module card gives you that in one click, and it is the single most
-useful thing you can attach.
+Something not working? Tell us. You do not need to diagnose it, describe what you saw, and hand us the state of the device it happened on. The **`{ }`** link on every module card gives you that in one click, and it is the single most useful thing you can attach.
 
 👉 **[Open an issue on GitHub](https://github.com/MoonModules/projectMM/issues)**
 
@@ -16,10 +13,10 @@ supposed to work like this?".
 ## Checklist
 
 - Search the [existing issues](https://github.com/MoonModules/projectMM/issues)
-  first — it may already be reported, or already fixed
+  first, it may already be reported, or already fixed
 - Check the Firmware card: are you on the current version?
 - For bugs, include the `{ }` output (below)
-- For feature requests, describe the use case — no template needed
+- For feature requests, describe the use case, no template needed
 
 ---
 
@@ -41,13 +38,11 @@ Actual behavior:
 
 ### 2. Attach the `{ }` output
 
-⚠️ **This is the most useful diagnostic data.** It is a JSON snapshot of the
-module's live state — its type, every control value, its status line and its
-timing — exactly as it was when things went wrong.
+⚠️ **This is the most useful diagnostic data.** It is a JSON snapshot of the module's live state, its type, every control value, its status line and its timing, exactly as it was when things went wrong.
 
 **How to get it:**
 
-1. Find the card that misbehaved — the effect that went dark, the driver that
+1. Find the card that misbehaved, the effect that went dark, the driver that
    isn't lighting, the network card that won't connect
 2. Click the **`{ }`** link in its title row, next to the `?`
 3. Select all in the new tab (Ctrl/Cmd+A), copy, and paste it into the issue
@@ -55,9 +50,9 @@ timing — exactly as it was when things went wrong.
 **Which cards to include:**
 
 1. **Always: System.** Names the chip, firmware variant, build, SDK version and
-   the last boot reason — how we tell a board problem from a firmware one
+   the last boot reason, how we tell a board problem from a firmware one
 2. **Always: Firmware.** The exact version and build you are running
-3. **The affected card itself** — Effects, Drivers, Layouts, Network, whichever
+3. **The affected card itself**, Effects, Drivers, Layouts, Network, whichever
    one went wrong
 4. **If it is about lights:** the Layer card, which carries the geometry
 5. **If it is about pins or wiring:** the Drivers card
@@ -80,32 +75,27 @@ API output:
 
 ### 3. Add a photo
 
-If it is something you can see — a fixture showing the wrong colours, a UI in a
-state that looks wrong — a photo or screenshot says it faster than a paragraph.
+If it is something you can see, a fixture showing the wrong colors, a UI in a state that looks wrong, a photo or screenshot says it faster than a paragraph.
 
 ---
 
 ## If the device crashed or won't boot
 
-The `{ }` link needs a device that still serves its interface. When it doesn't,
-tell us instead:
+The `{ }` link needs a device that still serves its interface. When it doesn't, tell us instead:
 
-- **What the LEDs did** — nothing at all, a brief flash, a repeating pattern
+- **What the LEDs did**: nothing at all, a brief flash, a repeating pattern
 - **What changed** just before it started, even if it seems unrelated: a setting,
   a script edit, a firmware update, a cable moved
 - **Does it repeat?** Power-cycle it. A device that boots once and fails the next
   time is a different problem from one that never boots
 - **The serial boot log**, if you have the device on USB and can run a monitor.
-  If you can't, say so — we will not ask you to set up a toolchain to file a bug
+  If you can't, say so, we will not ask you to set up a toolchain to file a bug
 
-One more thing worth checking yourself: `bootReason` on the System card. `PANIC`,
-`TASK_WDT` or `BROWNOUT` means the device crashed or lost power rather than
-merely misbehaving, and saying so in the first line of the report saves a round
-trip.
+One more thing worth checking yourself: `bootReason` on the System card. `PANIC`, `TASK_WDT` or `BROWNOUT` means the device crashed or lost power rather than merely misbehaving, and saying so in the first line of the report saves a round trip.
 
 ---
 
 ## Feature requests
 
 Describe the use case, what you expect it to do, and what you have tried instead.
-No template, no diagnostics — just tell us what you want to build.
+No template, no diagnostics, just tell us what you want to build.
diff --git a/docs/how-to/panel-cards.md b/docs/how-to/panel-cards.md
index 86939d37..fbc02209 100644
--- a/docs/how-to/panel-cards.md
+++ b/docs/how-to/panel-cards.md
@@ -55,9 +55,7 @@ The cards have no buffering and no flow control. They latch the image when the s
 
 The failure mode is the confusing part: **nothing errors**. The link is up, frames go out, and the panels tear, show wrong rows, or never latch. That is why projectMM reads the *negotiated* speed and warns you, rather than letting a slow link look like a format bug. It still sends, since a small wall on 100 Mbit is often fine, but if your picture is unstable, check this first.
 
-The cards can also be picky about negotiating a gigabit link with a 100 Mbit controller. A gigabit
-switch in between is the remedy: the card negotiates gigabit with the switch, the switch buffers, and
-the controller's slower link stops being the card's problem. This applies to the **P4 and the S3**.
+The cards can also be picky about negotiating a gigabit link with a 100 Mbit controller. A gigabit switch in between is the remedy: the card negotiates gigabit with the switch, the switch buffers, and the controller's slower link stops being the card's problem. This applies to the **P4 and the S3**.
 An **S31 is gigabit on its own** and connects directly.
 
 ---
@@ -72,23 +70,16 @@ An **S31 is gigabit on its own** and connects directly.
 | [7. Card firmware, and the flicker](#7-card-firmware-and-the-flicker) | The v13 defect, and the LED Upgrade 4.0 downgrade that clears it |
 | [8. When it does not light up](#8-when-it-does-not-light-up) | Symptom to cause |
 
-The two halves are independent: an ESP32 and a desktop drive the same card the same way, and neither
-is a prerequisite for the other. Pick whichever hardware you have.
+The two halves are independent: an ESP32 and a desktop drive the same card the same way, and neither is a prerequisite for the other. Pick whichever hardware you have.
 
 ---
 
 ## 4. Set the panels up in LED Vision
 
-The receiving card has to know what it is driving before projectMM sends it anything: how big each
-panel is, how many there are, and which driver IC they use. That configuration lives **on the card**,
-written once with ColorLight's own **[LEDVision](https://en.colorlightinside.com/product/download/380)**,
-and it is why projectMM itself needs no panel wiring settings at all (see
-[§5.4](#54-describe-the-wall)).
+The receiving card has to know what it is driving before projectMM sends it anything: how big each panel is, how many there are, and which driver IC they use. That configuration lives **on the card**, written once with ColorLight's own **[LEDVision](https://en.colorlightinside.com/product/download/380)**, and it is why projectMM itself needs no panel wiring settings at all (see [§5.4](#54-describe-the-wall)).
 
 **Which version.** An **8.x** build is what people running these cards in this scene actually use:
-this project's own wall is set up with **8.8**, and the walkthrough linked below uses **8.5**. Newer
-releases exist, and whether they are equally suitable here has not been established, so the safe
-advice is to take an 8.x build and only move if you have a reason to.
+this project's own wall is set up with **8.8**, and the walkthrough linked below uses **8.5**. Newer releases exist, and whether they are equally suitable here has not been established, so the safe advice is to take an 8.x build and only move if you have a reason to.
 
 > **Worth watching first:** [Setting up a Colorlight Card with FPP v6.3 and LED Vision 8.5](https://www.youtube.com/watch?v=L4lHbwUszAs)
 > walks through the whole card-and-panel setup on video. It drives the card from FPP rather than
@@ -103,17 +94,14 @@ advice is to take an 8.x build and only move if you have a reason to.
 2. Open LEDVision and go to the **receiving-card** setup (usually `Settings` then a receiving-card
    or `Screen` panel; some versions ask for a password, commonly `168`).
 3. Load the panel definition. Either pick your panel from the built-in module list, or load the
-   `.rcfgx` / `.rcvx` file the panel supplier provided, which is the reliable route for a panel that
-   is not a well-known model.
+   `.rcfgx` / `.rcvx` file the panel supplier provided, which is the reliable route for a panel that is not a well-known model.
 4. Set the **cabinet** size: how many pixels one card drives, across and down.
 5. Set the **panel arrangement**: how the HUB75 ribbons chain, and which physical panel is first.
    This is the step that makes the card, not projectMM, responsible for panel order.
 6. **Send to the receiving card**, then **Save** so the configuration survives a power cycle. Saving
-   is a separate action from sending in most versions, and skipping it is the usual reason a wall
-   comes back wrong after being unplugged.
+   is a separate action from sending in most versions, and skipping it is the usual reason a wall comes back wrong after being unplugged.
 
-When this is right, a test pattern from LEDVision fills the wall correctly. Get to that point before
-introducing projectMM: it separates "the panels are wired and configured" from "the sender works".
+When this is right, a test pattern from LEDVision fills the wall correctly. Get to that point before introducing projectMM: it separates "the panels are wired and configured" from "the sender works".
 
 ---
 
@@ -129,10 +117,7 @@ introducing projectMM: it separates "the panels are wired and configured" from "
 
 Other ESP32 variants do not ship panel-card support: their Ethernet is 100 Mbit at best, and most have none at all.
 
-The S3 needs a **W5500 Ethernet module** wired to its SPI pins, which the S3 firmwares already
-support; its pins come from the board entry in `deviceModels.json`. Because W5500 is 100 Mbit and
-sits behind SPI, it is the configuration most likely to need the gigabit switch described in
-[§2](#the-one-hardware-fact-that-decides-everything).
+The S3 needs a **W5500 Ethernet module** wired to its SPI pins, which the S3 firmwares already support; its pins come from the board entry in `deviceModels.json`. Because W5500 is 100 Mbit and sits behind SPI, it is the configuration most likely to need the gigabit switch described in [§2](#the-one-hardware-fact-that-decides-everything).
 
 ### 5.2 Flash it
 
@@ -153,47 +138,26 @@ Panel-card support is compiled in per firmware, and it is already on for the boa
 **On powering panels from the same supply as the board.** This works, and plenty of small setups run
 that way. What it costs you is headroom.
 
-HUB75 panels draw far less than a naive count suggests, and the reason is **multiplexing**: the panel
-lights one group of scan rows at a time, cycling fast enough to look continuous, so at 1/16 or 1/32
-scan only a fraction of the LEDs are on at any instant. This is why these panels have a reputation
-for modest consumption. Size the supply from the rating on your panel's own datasheet rather than
-from pixels multiplied by LED current.
+HUB75 panels draw far less than a naive count suggests, and the reason is **multiplexing**: the panel lights one group of scan rows at a time, cycling fast enough to look continuous, so at 1/16 or 1/32 scan only a fraction of the LEDs are on at any instant. This is why these panels have a reputation for modest consumption. Size the supply from the rating on your panel's own datasheet rather than from pixels multiplied by LED current.
 
-The failure when you do run short is not a clean one: the 5 V rail sags, and a sagging rail shows up
-as flicker, color shifts, or the controller resetting mid-frame. None of those look like a power
-problem, which is why they cost an evening. A wall that is stable at 30% brightness and misbehaves at
-100% is telling you this is the problem, not the network.
+The failure when you do run short is not a clean one: the 5 V rail sags, and a sagging rail shows up as flicker, color shifts, or the controller resetting mid-frame. None of those look like a power problem, which is why they cost an evening. A wall that is stable at 30% brightness and misbehaves at 100% is telling you this is the problem, not the network.
 
 Nothing needs an IP address: the panel link is below IP entirely.
 
 **That is also how you reach the UI.** The board's Ethernet port is now carrying panel data, so
-WiFi is what serves the web interface. Leave WiFi configured as normal; it is unrelated to the panel
-link and the two do not interfere. On the P4 this decides which firmware to flash, because the
-`-eth` variant has no WiFi compiled in at all: use `-eth-wifi`. The S31 and S3 firmwares carry both
-already.
+WiFi is what serves the web interface. Leave WiFi configured as normal; it is unrelated to the panel link and the two do not interfere. On the P4 this decides which firmware to flash, because the `-eth` variant has no WiFi compiled in at all: use `-eth-wifi`. The S31 and S3 firmwares carry both already.
 
 ### 5.4 Describe the wall
 
 The driver has **no geometry controls**. The wall's shape lives in the Layout, once, so that everything else (effects, modifiers, the preview) sees the same picture.
 
 **A plain Grid is usually all you need.** Two 128x64 panels stacked is a 128x128 grid, and that is
-the whole configuration. The reason it is that simple is worth knowing: the driver reads only the
-wall's width and height and sends the image row by row. Which physical panel a row lands on, and in
-what order the HUB75 ribbons chain, was already settled on the card in
-[§4](#4-set-the-panels-up-in-led-vision). The card owns panel arrangement; projectMM owns the
-picture.
-
-That is also why this needs none of the physical detail you may have filled in elsewhere. An output
-page that asks for scan rate, address lines and chain order is describing panels driven *directly*,
-where the software has to generate the HUB75 timing itself. Through a receiving card, none of that is
-the sender's business: the card generates the timing, and the sender hands it an image. That holds
-for any sender, [FPP](https://github.com/FalconChristmas/fpp) included, which reaches these cards
-over Ethernet exactly as projectMM does.
+the whole configuration. The reason it is that simple is worth knowing: the driver reads only the wall's width and height and sends the image row by row. Which physical panel a row lands on, and in what order the HUB75 ribbons chain, was already settled on the card in [§4](#4-set-the-panels-up-in-led-vision). The card owns panel arrangement; projectMM owns the picture.
+
+That is also why this needs none of the physical detail you may have filled in elsewhere. An output page that asks for scan rate, address lines and chain order is describing panels driven *directly*, where the software has to generate the HUB75 timing itself. Through a receiving card, none of that is the sender's business: the card generates the timing, and the sender hands it an image. That holds for any sender, [FPP](https://github.com/FalconChristmas/fpp) included, which reaches these cards over Ethernet exactly as projectMM does.
 
 **When you need the Panels layout instead.** It exists for walls where projectMM, not a card, owns
-the ordering: addressable panels wired as one long pixel strip, where the strip snakes from panel to
-panel and the layout has to undo that. Its controls are about **wiring order**, which a HUB75 ribbon
-does not have.
+the ordering: addressable panels wired as one long pixel strip, where the strip snakes from panel to panel and the layout has to undo that. Its controls are about **wiring order**, which a HUB75 ribbon does not have.
 
 | Control | Meaning |
 |---|---|
@@ -203,10 +167,7 @@ does not have.
 | `wiringOrderP`, `X++P`, `Y++P`, `snakeP` | How the panels themselves are ordered |
 
 **Do HUB75 panels snake?** Not in the sense these controls mean. A HUB75 panel is addressed by row
-and column over the ribbon, so its internal pixel order is fixed by the panel's own driver ICs and
-is not something a layout re-maps. A *chain* of panels can be arranged in any order, including a
-serpentine one, but that is configured on the card, not here. So if you are driving panels through a
-receiving card and every other row looks reversed, the setting to revisit is in LEDVision.
+and column over the ribbon, so its internal pixel order is fixed by the panel's own driver ICs and is not something a layout re-maps. A *chain* of panels can be arranged in any order, including a serpentine one, but that is configured on the card, not here. So if you are driving panels through a receiving card and every other row looks reversed, the setting to revisit is in LEDVision.
 
 ### 5.5 Add the driver
 
@@ -306,16 +267,13 @@ The second row is worth taking seriously, because projectMM sends a full frame e
 ### Reading and changing the version
 
 **Use [LEDUpgrade](https://en.colorlightinside.com/product/download/383) 4.0 and firmware 11.09.**
-That is the proven combination, and the easiest one, because 11.09 ships inside LEDUpgrade 4.0: it is
-in the preset list, so there is no firmware file to find. Treat any other pairing as a detour.
+That is the proven combination, and the easiest one, because 11.09 ships inside LEDUpgrade 4.0: it is in the preset list, so there is no firmware file to find. Treat any other pairing as a detour.
 
 **Why not 5.0.** Version 5.0 ships no pre-v12 firmware at all, so it cannot do this downgrade from
 its preset list however long you fight it.
 
 **Why 11.09 rather than something older.** Anything before v12 clears the flicker, but older is not
-automatically safer: cards on 11.08 were reported strobing white, which 11.09 fixes. 11.09 is the
-newest build on the safe side of the defect, so it carries the most fixes while carrying none of the
-flicker.
+automatically safer: cards on 11.08 were reported strobing white, which 11.09 fixes. 11.09 is the newest build on the safe side of the defect, so it carries the most fixes while carrying none of the flicker.
 
 > **On Windows, if LEDUpgrade cannot find the card.** A built-in Ethernet port can be held by
 > something else in the stack, and Hyper-V's virtual switch is the usual culprit: it binds the
@@ -341,8 +299,7 @@ Set the driver's `firmware` control to match the card:
 | `v12 and older` | **The default**, and a downgraded card. Brightness and sync go out once. |
 | `v13 and newer` | A stock card. Both go out twice, which is the copy this firmware acts on. |
 
-The default is the downgraded generation on purpose: this page's own advice is to move a v13 card
-off it, so the setting is already right when you finish rather than being one last unexplained step.
+The default is the downgraded generation on purpose: this page's own advice is to move a v13 card off it, so the setting is already right when you finish rather than being one last unexplained step.
 
 The mismatch is not subtle in one direction: leave a downgraded card on `v13 and newer` and it receives a second sync, treats it as another latch, aborts the refresh already running, and the wall updates once every few seconds.
 
diff --git a/docs/index.md b/docs/index.md
index 51e5663d..46aede99 100644
--- a/docs/index.md
+++ b/docs/index.md
@@ -2,9 +2,7 @@
 
 High-performance LED &amp; DMX lighting control for ESP32 and beyond.
 
-[:material-flash: Flash an ESP32 from your browser](/projectMM/install/){ .md-button .md-button--primary }
-&nbsp;
-[:material-github: GitHub](https://github.com/MoonModules/projectMM){ .md-button }
+[:material-flash: Flash an ESP32 from your browser](/projectMM/install/){ .md-button .md-button--primary } &nbsp; [:material-github: GitHub](https://github.com/MoonModules/projectMM){ .md-button }
 
 !!! tip "New here?"
     The [Getting started](gettingstarted.md) guide walks you from a blank ESP32 to your first running light show, step by step, with no build tools required.
diff --git a/docs/moonmodules/core/services.md b/docs/moonmodules/core/services.md
index 25b18e19..11565f0e 100644
--- a/docs/moonmodules/core/services.md
+++ b/docs/moonmodules/core/services.md
@@ -47,8 +47,7 @@ Detail: [technical](moxygen/AudioService.md) · [the sync packet](../light/moxyg
 Receives [OSC](https://opensoundcontrol.stanford.edu/) over UDP and writes it onto this device's controls, so a fader in Resolume, TouchDesigner or TouchOSC drives projectMM directly. It owns no surface of its own: everything lands in the same control writes the HTTP API and the UI use, so every validator still runs. Addresses, feedback and setup: ⌄ details.
 
 - `listen` — receive OSC (default **off**). This opens an unauthenticated UDP port that writes
-  controls, on the same LAN-trust basis as the Art-Net and audio-sync receivers, so it is a
-  capability you turn on rather than one every device carries.
+  controls, on the same LAN-trust basis as the Art-Net and audio-sync receivers, so it is a capability you turn on rather than one every device carries.
 - `port` — the UDP port (default 9000, what TouchOSC uses). Applies live.
 - `status` — listening, off, or why the port could not be opened.
 
@@ -111,8 +110,7 @@ A Service added per board: **a MoonLive script that reads hardware and drives co
 Why a script rather than another module: a mapping row is right for a button and wrong for anything with a condition in it. A script holds state and chooses between outcomes.
 
 - `script`: which `.mls` file to run, picked from the script library. Naming a different one
-  recompiles live; a compile error shows on the status line and the service does nothing until it is
-  fixed.
+  recompiles live; a compile error shows on the status line and the service does nothing until it is fixed.
 - Everything the script declares appears as a real control on the card, so a slider move lands
   without a recompile.
 
@@ -159,21 +157,12 @@ A desktop device is picked by list position, so re-pick if the OS reorders them;
 
 #### WLED audio sync: what is on the wire
 
-Sending and receiving both use the **multicast address 239.0.0.1**, which is what WLED's own
-usermod does (`beginMulticast` on both ends). It never uses broadcast, so a broadcast sender is
-inaudible to WLED and a receiver that only binds the port never hears WLED. This is a
-network-layer address, unrelated to any device grouping.
+Sending and receiving both use the **multicast address 239.0.0.1**, which is what WLED's own usermod does (`beginMulticast` on both ends). It never uses broadcast, so a broadcast sender is inaudible to WLED and a receiver that only binds the port never hears WLED. This is a network-layer address, unrelated to any device grouping.
 
 **Port 11988 is the WLED contract**, and `syncPort` defaults to it. The port is configurable for
-projectMM peers that want a private stream, but a custom port is no longer WLED-compatible: the
-endpoint WLED speaks is 239.0.0.1:11988 specifically.
+projectMM peers that want a private stream, but a custom port is no longer WLED-compatible: the endpoint WLED speaks is 239.0.0.1:11988 specifically.
 
-Multicast is also the better neighbour, with a caveat worth knowing: a switch or access point that
-does **IGMP snooping** forwards the group only to the ports that joined it, so the other hosts
-never see the traffic at all. Without snooping the switch floods it exactly like broadcast, and on
-WiFi it goes out at the lowest basic rate to every station. So multicast can reduce how many hosts
-have to process ~40 packets a second, but it does not guarantee it. See
-[multicast and IGMP snooping](../../explanation/architecture/moonlight.md#multicast-and-igmp-snooping).
+Multicast is also the better neighbour, with a caveat worth knowing: a switch or access point that does **IGMP snooping** forwards the group only to the ports that joined it, so the other hosts never see the traffic at all. Without snooping the switch floods it exactly like broadcast, and on WiFi it goes out at the lowest basic rate to every station. So multicast can reduce how many hosts have to process ~40 packets a second, but it does not guarantee it. See [multicast and IGMP snooping](../../explanation/architecture/moonlight.md#multicast-and-igmp-snooping).
 
 The 44-byte v2 packet is byte-compatible with WLED, with one field that is not yet equivalent:
 
@@ -187,43 +176,24 @@ The 44-byte v2 packet is byte-compatible with WLED, with one field that is not y
 | `FFT_Magnitude` | 0..255 internally, x16 on the wire | ~0..4096 | compatible |
 
 **The magnitude scale differs, so it is converted at the wire.** WLED sends the raw magnitude of
-its FFT's dominant bin, scaled so that "the end result is linear and ~4096 max" (its own comment
-where it divides the input samples by 16). Its effects then divide that by 4, 8 or 16 depending on
-the effect and treat the result as a byte, which is why their thresholds read `< 48` squelch and
-`> 144` full brightness. projectMM byte-scales the peak magnitude to 0..255 instead, through the
-same noise floor and gain conditioning as the 16 bands, so one pair of knobs governs the whole
-spectrum.
+its FFT's dominant bin, scaled so that "the end result is linear and ~4096 max" (its own comment where it divides the input samples by 16). Its effects then divide that by 4, 8 or 16 depending on the effect and treat the result as a byte, which is why their thresholds read `< 48` squelch and `> 144` full brightness. projectMM byte-scales the peak magnitude to 0..255 instead, through the same noise floor and gain conditioning as the 16 bands, so one pair of knobs governs the whole spectrum.
 
 projectMM keeps its own units internally and multiplies by 16 on send, dividing by 16 on receive.
-The factor is exact rather than a fudge: it is the divisor WLED's effects apply, so our full-scale
-255 arrives as 4080, right on WLED's own ~4096 design target, and every effect's thresholds land
-where they were tuned to. Adopting WLED's range internally was the alternative, and was rejected
-because that range is an artifact of FFT size and input scaling rather than a specification (WLED's
-own fallback path admits "no idea if 10000 is a good value"), and importing it would cost the
-property that one floor/gain pair conditions every value the service publishes, in exchange for
-resolution the receiving effects discard anyway when they divide back down to a byte.
+The factor is exact rather than a fudge: it is the divisor WLED's effects apply, so our full-scale 255 arrives as 4080, right on WLED's own ~4096 design target, and every effect's thresholds land where they were tuned to. Adopting WLED's range internally was the alternative, and was rejected because that range is an artifact of FFT size and input scaling rather than a specification (WLED's own fallback path admits "no idea if 10000 is a good value"), and importing it would cost the property that one floor/gain pair conditions every value the service publishes, in exchange for resolution the receiving effects discard anyway when they divide back down to a byte.
 
-A received magnitude is clamped to 255, since a real WLED source reaches ~9500 and an unclamped
-value would drive effects harder than locally analyzed audio ever could.
+A received magnitude is clamped to 255, since a real WLED source reaches ~9500 and an unclamped value would drive effects harder than locally analyzed audio ever could.
 
 Prior art: the WLED-MM audio-reactive usermod by **Frank ([@softhack007](https://github.com/softhack007))**, the most-used open-source audio-reactive LED implementation, whose adaptive noise-gate concept the analysis here descends from (analyzed with his permission); and **[@troyhacks](https://github.com/troyhacks/WLED)**, who reworked that DSP onto Espressif's [esp-dsp](https://github.com/espressif/esp-dsp) FFT, the same choice this service makes. The line-in path exists because **wladi ([myhome-control](https://shop.myhome-control.de))** supplied the hardware and pinout for the [MHC-WLED ESP32-P4 shield](../../reference/hardware/mhc-wled-esp32-p4-shield.md): its onboard PCM1808 I2S ADC is what `mclkPin` is for.
 
 ## OSC, details
 **Feedback: the device answers.** With `feedback` on, a control that changes anywhere (the web UI, a
-preset recall, an audio-reactive effect) is mirrored back to the surface, which is what keeps a
-client honest and what moves a motorised fader. `feedbackTo` names the receiver, or is left empty to
-answer whoever last wrote to us; `feedbackPort` is where that client LISTENS, which is not the port
-we listen on (Open Stage Control calls its own `osc-port`).
-
-A client learns the current state three ways: when it first writes to us from a new address, when
-its address changes, and whenever it sends **`/mm/hello`**. The last one exists because a client
-restarting on the SAME address is invisible to the other two, and most controllers send nothing of
-their own on load, so every widget would show its layout file's defaults until the user moved one.
+preset recall, an audio-reactive effect) is mirrored back to the surface, which is what keeps a client honest and what moves a motorised fader. `feedbackTo` names the receiver, or is left empty to answer whoever last wrote to us; `feedbackPort` is where that client LISTENS, which is not the port we listen on (Open Stage Control calls its own `osc-port`).
+
+A client learns the current state three ways: when it first writes to us from a new address, when its address changes, and whenever it sends **`/mm/hello`**. The last one exists because a client restarting on the SAME address is invisible to the other two, and most controllers send nothing of their own on load, so every widget would show its layout file's defaults until the user moved one.
 The shipped session has a `sync from device` button for exactly this.
 
 **Setting one up**, from installing the app to using it from a phone, is its own page:
-[Driving projectMM from a phone or tablet](../../how-to/control-surface.md). It needs no
-checkout and no tooling, just the app and the session file from the latest release.
+[Driving projectMM from a phone or tablet](../../how-to/control-surface.md). It needs no checkout and no tooling, just the app and the session file from the latest release.
 
 **Addresses.** These are a public contract: a TouchOSC layout built against them keeps working, so
 they stay small and boring.
@@ -236,9 +206,7 @@ they stay small and boring.
 | `/mm/hello` | anything, or nothing | resend every value to the sender |
 | `/mm/control/<Module>/<control>` | float 0..1 or int 0..255 | any control directly |
 
-Both argument forms are accepted because controllers disagree: apps send a float in 0..1, hardware
-bridges send an int in the target's range. Out-of-range values are clamped rather than ignored, so
-a controller sending 0..127 does something sensible instead of appearing dead.
+Both argument forms are accepted because controllers disagree: apps send a float in 0..1, hardware bridges send an int in the target's range. Out-of-range values are clamped rather than ignored, so a controller sending 0..127 does something sensible instead of appearing dead.
 
 Send one from the bench with `uv run moondeck/check/send_osc.py <ip> /mm/fader/1 0.75`.
 
@@ -252,15 +220,9 @@ uv run moondeck/run/run_open_stage_control.py                   # device on this
 uv run moondeck/run/run_open_stage_control.py --host 192.168.1.42
 ```
 
-Open **http://127.0.0.1:8088** and the surface is there. On the device, turn `listen` and `feedback`
-on; nothing else needs configuring, because the launcher passes the session, the send address and
-the listen port as arguments rather than leaving them to be typed into a settings panel. The session
-itself sends `/mm/hello` when the page loads, so every widget shows the device's real values straight
-away instead of its layout file's defaults, and a browser refresh re-reads them.
+Open **http://127.0.0.1:8088** and the surface is there. On the device, turn `listen` and `feedback` on; nothing else needs configuring, because the launcher passes the session, the send address and the listen port as arguments rather than leaving them to be typed into a settings panel. The session itself sends `/mm/hello` when the page loads, so every widget shows the device's real values straight away instead of its layout file's defaults, and a browser refresh re-reads them.
 
-It runs **headless**: a web server rather than a desktop window. That is deliberate. The surface is
-then reachable from a phone or another laptop on the same network (the launcher prints those URLs),
-and on macOS it sidesteps the quarantine dialog an unsigned download otherwise raises.
+It runs **headless**: a web server rather than a desktop window. That is deliberate. The surface is then reachable from a phone or another laptop on the same network (the launcher prints those URLs), and on macOS it sidesteps the quarantine dialog an unsigned download otherwise raises.
 
 | | |
 |---|---|
@@ -270,30 +232,19 @@ and on macOS it sidesteps the quarantine dialog an unsigned download otherwise r
 | `--app` | the Open Stage Control binary, when it is not on PATH or in the usual place |
 | `--gui` | also open the desktop window; by default it is the server alone |
 
-The launcher looks on PATH first, then in each platform's default install location. **Windows and
-Linux are untested**: the paths are the ones those installers use, but only macOS has been run. If
-it cannot find the app, `--app` takes the full path and that always works.
+The launcher looks on PATH first, then in each platform's default install location. **Windows and Linux are untested**: the paths are the ones those installers use, but only macOS has been run. If it cannot find the app, `--app` takes the full path and that always works.
 
 **A ready-made control surface.** A session of the switches, encoders and faders ships as a release
-asset (`projectMM-control-surface.json`) and lives in the repo at
-[`docs/reference/examples/open-stage-control.json`](../../reference/examples/open-stage-control.json).
+asset (`projectMM-control-surface.json`) and lives in the repo at [`docs/reference/examples/open-stage-control.json`](../../reference/examples/open-stage-control.json).
 Editing the layout needs `read-only` off in the launcher.
 
 <img src="../../assets/core/OscModule-open-stage-control.png" width="600" alt="The shipped Open Stage Control session beside projectMM's own Control card: eight switches, eight encoders and eight faders in both">
 
 Driving the device from that session, beside the Control card it mirrors.
 
-It binds only to `/mm/switch/N`, `/mm/encoder/N` and `/mm/fader/N`, N being 1 to 8, on purpose. A surface should
-address the SURFACE, and [Control](system.md#control) decides what each one drives, so one layout keeps
-working as assignments change and a hardware desk lands on the same bindings. Reaching past it to
-`/mm/control/<Module>/<control>` also works and is the right answer for a one-off, but it hard-codes
-into the layout a mapping that belongs on the device. Two have targets today, `switch1`
-(`Drivers.on`) and `fader1` (`Drivers.brightness`); the rest wait for a target picker.
+It binds only to `/mm/switch/N`, `/mm/encoder/N` and `/mm/fader/N`, N being 1 to 8, on purpose. A surface should address the SURFACE, and [Control](system.md#control) decides what each one drives, so one layout keeps working as assignments change and a hardware desk lands on the same bindings. Reaching past it to `/mm/control/<Module>/<control>` also works and is the right answer for a one-off, but it hard-codes into the layout a mapping that belongs on the device. Two have targets today, `switch1` (`Drivers.on`) and `fader1` (`Drivers.brightness`); the rest wait for a target picker.
 
-The session also carries a **pad grid**, and those pads are inert: `/mm/pad/N` has no route in the
-OSC module yet, so pressing one sends a message nothing reads. It ships anyway because the grid is
-the layout a preset launcher wants and the addresses are the ones it will use; treat it as a
-placeholder rather than as part of the contract above.
+The session also carries a **pad grid**, and those pads are inert: `/mm/pad/N` has no route in the OSC module yet, so pressing one sends a message nothing reads. It ships anyway because the grid is the layout a preset launcher wants and the addresses are the ones it will use; treat it as a placeholder rather than as part of the contract above.
 
 **It does not reach a Mackie desk.** The X-Touch and QCon Pro G2 speak Mackie Control over MIDI,
 not OSC: see [control surfaces](../../reference/hardware/control-surfaces.md) for what would.
@@ -301,16 +252,9 @@ not OSC: see [control surfaces](../../reference/hardware/control-surfaces.md) fo
 ## Infrared, details
 Set a row's `learn` and the next code received binds to it, which is how any remote works without a shipped code table. Arming one row disarms any other, so a code cannot bind twice. A fresh service starts with no rows: add one, learn a key, pick a target. The status line reports whether the channel actually opened, not merely that a pin is set; on some boards the receiver shares its pin with another peripheral through a board switch.
 
-Nothing is fixed in firmware. A row IS the binding: learn a key onto it, pick what it drives from
-the target dropdown, and pick whether the press toggles that control, or nudges it by a value. A
-handset with twenty keys is twenty rows. `set` is offered only where an input reports a release, so
-it is unavailable here: a remote code is a single event, and a `set` row would latch the control
-with nothing able to clear it.
+Nothing is fixed in firmware. A row IS the binding: learn a key onto it, pick what it drives from the target dropdown, and pick whether the press toggles that control, or nudges it by a value. A handset with twenty keys is twenty rows. `set` is offered only where an input reports a release, so it is unavailable here: a remote code is a single event, and a `set` row would latch the control with nothing able to clear it.
 
-One key binds to one row. Learning a key that another row already holds moves the binding rather
-than duplicating it, because dispatch fires the first row holding a code and a duplicate could
-never run.
+One key binds to one row. Learning a key that another row already holds moves the binding rather than duplicating it, because dispatch fires the first row holding a code and a duplicate could never run.
 
-The status line reports setup state ("set pin to receive" / "ready"), the learn prompt, a binding
-("learned 0x..."), what a press did or why it did not, and an unbound code ("received 0x...
+The status line reports setup state ("set pin to receive" / "ready"), the learn prompt, a binding ("learned 0x..."), what a press did or why it did not, and an unbound code ("received 0x...
 (unassigned)").
diff --git a/docs/moonmodules/core/system.md b/docs/moonmodules/core/system.md
index 2675647f..b804e01a 100644
--- a/docs/moonmodules/core/system.md
+++ b/docs/moonmodules/core/system.md
@@ -8,7 +8,7 @@ The device's fixed infrastructure: identity, network, provisioning, firmware, an
 
 ### System
 
-The device's identity and vitals: name (behind mDNS `<name>.local`, the SoftAP SSID, the DHCP hostname), uptime, heap, and per-module footprint reporting. Its fixed inspection children (Tasks, I2C scan) hang beneath it.
+The device's identity and vitals: name (behind mDNS `<name>.local`, the SoftAP SSID, the DHCP hostname), uptime, heap and the tick it renders at. Its fixed inspection children (Tasks, I2C scan) hang beneath it.
 
 <img src="../../assets/core/SystemModule.png" width="300" alt="System module controls">
 
@@ -16,7 +16,9 @@ The device's identity and vitals: name (behind mDNS `<name>.local`, the SoftAP S
 - `deviceModel`: the board model (drives the installer catalog entry).
 - `mode`: how much of the UI is shown, `user`, `expert` (🎚️) or `developer` (🔧), cumulative.
 - `logLevel`: serial verbosity, defaulting to Warn. The first 60 s always logs at Info.
-- read-only vitals: `uptime`, `fps`, `heap`, `psram`, `flash`, `chip`, and per-module footprint.
+- read-only rate: `uptime`, `fps`, `tickTimeUs`.
+- read-only memory: `heap`, `psram`, `maxBlock`, `flash`.
+- read-only identity: `mac`, `chip`, `cpu`, `sdk`, `bootReason`.
 
 Detail: [technical](moxygen/SystemModule.md)
 
@@ -30,12 +32,12 @@ WiFi / Ethernet connectivity, static-IP configuration, RSSI and TX-power reporti
 
 <img src="../../assets/core/NetworkModule.png" width="300" alt="Network module controls">
 
-- `mode`: WiFi / Ethernet / off.
 - `ssid` / `password`: WiFi credentials.
 - `mDNS`: the `<name>.local` hostname.
-- `addressing`: DHCP or static; static exposes IP / gateway / subnet / DNS fields.
-- `ethType` / `ethPhyAddr` / `ethRstGpio` / …, Ethernet PHY configuration.
-- read-only: `rssi` (dBm), `txPower` (dBm).
+- `addressing`: DHCP or static; static reveals `ip`, `gateway`, `subnet` and `dns`.
+- `ethBoard`: the board's Ethernet wiring by name; `Custom` exposes every pin.
+- `txPowerSetting`: caps the radio's transmit power for a board that browns out; 0 lifts it.
+- read-only: `mode`, the interface in use, with `rssi` / `txPower` (dBm) on a live radio.
 
 Detail: [technical](moxygen/NetworkModule.md)
 
diff --git a/docs/moonmodules/core/ui.md b/docs/moonmodules/core/ui.md
index 79083478..40f44cf7 100644
--- a/docs/moonmodules/core/ui.md
+++ b/docs/moonmodules/core/ui.md
@@ -1,19 +1,12 @@
 # Web UI
 
-The projectMM web UI as shipped, the render layer over the MoonModule tree. This page is the UI's
-own implementation spec (status bar, cards, control rendering, styling, the no-rebuild update
-contract). The **high-level architecture**, hand-maintained files, MoonModule-driven rendering, the
-light-domain plug-in points, lives in [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui);
-the **backend contract** it consumes (every `/api/*` endpoint, the `/ws` frame shape, the control
-descriptors) is owned by [HttpServerModule](moxygen/HttpServerModule.md); the **emoji legend** the
-cards and picker render is [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend).
+The projectMM web UI as shipped, the render layer over the MoonModule tree. This page is the UI's own implementation spec (status bar, cards, control rendering, styling, the no-rebuild update contract). The **high-level architecture**, hand-maintained files, MoonModule-driven rendering, the light-domain plug-in points, lives in [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui); the **backend contract** it consumes (every `/api/*` endpoint, the `/ws` frame shape, the control descriptors) is owned by [HttpServerModule](moxygen/HttpServerModule.md); the **emoji legend** the cards and picker render is [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend).
 This page covers only what those don't: the browser-side rendering behavior.
 
 ## Interaction principles
 
 - **Two timescales for inputs.** Local UI feedback (a slider's value label, a toggle's state)
-  updates within ~20 ms of the input event; network sends are debounced (150 ms slider, 500 ms text),
-  so a slider feels instant even when the server is busy.
+  updates within ~20 ms of the input event; network sends are debounced (150 ms slider, 500 ms text), so a slider feels instant even when the server is busy.
 - **No write-back races.** A `dragTs[moduleId:key]` cooldown of 1 s stops an incoming WS push from
   overwriting a control the user is actively manipulating.
 - **No DOM rebuilds on state updates.** The initial render builds the tree; state updates patch values
@@ -24,7 +17,7 @@ This page covers only what those don't: the browser-side rendering behavior.
 ```text
 ┌─────────────────────────────────────────────────────────────┐
 │  Status bar: ☰ · logo · brand · device name · sys stats     │
-│              · WS dot · reconnect · reboot · theme toggle   │
+│              · WS dot · reboot · theme toggle               │
 ├──────────────┬──────────────────────────────────────────────┤
 │  Side nav    │  Sticky preview canvas (shrinks on scroll)   │
 │  (root list) ├──────────────────────────────────────────────┤
@@ -55,11 +48,9 @@ Fixed top, 44 px, left-to-right:
 5. **System stats**, `uptime · NN K free` (uptime `Xd Yh Zm Ws`, free heap KB), read from the
    SystemModule controls in the WS state push; no separate polling endpoint.
 6. **WebSocket dot**, green = connected, gray = disconnected. The socket auto-reconnects with
-   exponential backoff (500 ms → 5 s) on close; no manual reconnect button (a page reload covers the
-   rare half-dead socket).
+   exponential backoff (500 ms → 5 s) on close, and a page reload covers the rare half-dead socket.
 7. **Reboot button (⏻)**, red border via `data-crashed="true"` when `SystemModule.bootReason` is
-   PANIC / INT_WDT / TASK_WDT / BROWNOUT. Press-twice to confirm: the first click arms it (solid red),
-   a second sends `POST /api/reboot`. Disarms after 3 s or on pointer-leave, no `confirm()` popup.
+   PANIC / INT_WDT / TASK_WDT / BROWNOUT. Press-twice to confirm: the first click arms it (solid red), a second sends `POST /api/reboot`. Disarms after 3 s or on pointer-leave, no `confirm()` popup.
 8. **Theme toggle (☀/🌙)**, flips `[data-theme]` on `<body>`; preference in `localStorage['mm_theme']`.
 
 ## Side navigation
@@ -67,33 +58,20 @@ Fixed top, 44 px, left-to-right:
 A left column listing the root modules, one entry per top-level MoonModule.
 
 - **One root visible at a time.** Clicking a nav entry sets the selected root; `renderCards()` renders
-  only that root's card subtree. The selection persists in `localStorage['mm_selectedRoot']`; the
-  active entry is highlighted.
+  only that root's card subtree. The selection persists in `localStorage['mm_selectedRoot']`; the active entry is highlighted.
 - **Hamburger toggle.** ☰ toggles `body.nav-open`. On wide screens (≥ 820 px) the nav is a static
-  column the hamburger collapses/expands; on narrow screens (< 820 px) it's a slide-in drawer over a
-  dimming overlay (click the overlay or press Esc to close).
+  column the hamburger collapses/expands; on narrow screens (< 820 px) it's a slide-in drawer over a dimming overlay (click the overlay or press Esc to close).
 - **The nav states its own order** (`NAV_ORDER` in app.js), which is deliberately NOT the order the
-  roots run in. `main.cpp` orders by dependency, Filesystem before anything that writes a file,
-  System before the modules that read its identity, and that is load-bearing, so it cannot be
-  reshuffled to suit a menu. It also reads as an implementation detail to a user: it puts System
-  first and the lights last. The nav instead groups by what someone is looking for: **Control**,
-  then the light pipeline in pipeline order (**Layouts → Effects → Drivers**), then the device
-  (**System, File Manager, Network, Services, Firmware**). A root the list does not name still
-  appears, after them, in scheduler order, adding a module never makes it invisible.
+  roots run in. `main.cpp` orders by dependency, Filesystem before anything that writes a file, System before the modules that read its identity, and that is load-bearing, so it cannot be reshuffled to suit a menu. It also reads as an implementation detail to a user: it puts System first and the lights last. The nav instead groups by what someone is looking for: **Control**, then the light pipeline in pipeline order (**Layouts → Effects → Drivers**), then the device (**System, File Manager, Network, Services, Firmware**). A root the list does not name still appears, after them, in scheduler order, adding a module never makes it invisible.
 - **No root reorder.** The order is fixed in that list; the nav does not drag-reorder.
 - **Footer** pinned to the bottom of the nav: social links (GitHub, Discord, Reddit, YouTube, inline
   SVG) and a `© <year> MoonLight` line.
 
-The WS state push carries the full module tree; only the selected root is rendered, and
-`updateValues()` patches just the visible cards (non-rendered roots have no DOM nodes, so their data
-is silently ignored). Per-root server-side filtering was evaluated and deferred, the 1 Hz push is
-cheap and the JSON is built through a streaming sink, so tree size is not a buffer-limit concern.
+The WS state push carries the whole module tree; only the selected root is rendered, and `updateValues()` patches just the visible cards (non-rendered roots have no DOM nodes, so their data is silently ignored). Filtering per root server-side buys little: the full tree arrives once on connect and each second after that carries only what changed, built through a streaming sink, so tree size is not a buffer-limit concern.
 
 ## Module card
 
-Each MoonModule renders as a card, with **child cards nested inside their parent card's box**, the
-parent's border encloses its children, so the tree shape is visible structurally, not just by
-indentation. Nesting depth shows as progressively lighter backgrounds and a left-margin indent.
+Each MoonModule renders as a card, with **child cards nested inside their parent card's box**, the parent's border encloses its children, so the tree shape is visible structurally, not just by indentation. Nesting depth shows as progressively lighter backgrounds and a left-margin indent.
 
 ```text
 ┌─ card ──────────────────────────────────┐
@@ -107,44 +85,28 @@ indentation. Nesting depth shows as progressively lighter backgrounds and a left
 ```
 
 - The parent's own controls render **above** its children; `+ add child` renders **below** them.
-  Child cards live in a `.card-children` wrapper appended into the parent card's DOM node (not flat
-  siblings); `renderModuleTree` recurses into the parent card, not into `main`.
+  Child cards live in a `.card-children` wrapper appended into the parent card's DOM node (not flat siblings); `renderModuleTree` recurses into the parent card, not into `main`.
 - **`{ }`** opens `GET /api/modules/{name}` in a new tab, that one module's live JSON, for issue
-  reports (see [Log an issue](../../how-to/logging-an-issue.md)). On EVERY card, unlike `✎`/`×` (user-editable
-  children only) and `?` (types that have a doc page).
+  reports (see [Log an issue](../../how-to/logging-an-issue.md)). On EVERY card, unlike `✎`/`×` (user-editable children only) and `?` (types that have a doc page).
 - **Enabled toggle** in the right-hand action cluster mirrors `MoonModule::enabled()`, a styled
-  `<button>` (transparent + muted border, 26×26); state shown by the glyph alone (accent **✓** on,
-  muted **○** off; `data-checked` + `aria-pressed`). Toggling fires `onEnabled()`; the Scheduler skips
-  disabled modules whose `respectsEnabled()` returns true (default).
+  `<button>` (transparent + muted border, 26×26); state shown by the glyph alone (accent **✓** on, muted **○** off; `data-checked` + `aria-pressed`). Toggling fires `onEnabled()`; the Scheduler skips disabled modules whose `respectsEnabled()` returns true (default).
 - **Emoji tags** next to the name show the same set the type picker uses: role + dimensional (both
-  UI-derived) + the curated `tags()` from `/api/types`. Identical identity across card and picker; the
-  assignments are [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend).
+  UI-derived) + the curated `tags()` from `/api/types`. Identical identity across card and picker; the assignments are [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend).
 - **Help link (?)** at the far right of the title row opens the module's doc page in a new tab. The
-  path comes from `docPath` in `/api/types` (engine-provided, relative to `docs/moonmodules/`); omitted
-  when the type declares none.
+  path comes from `docPath` in `/api/types` (engine-provided, relative to `docs/moonmodules/`); omitted when the type declares none.
 - **Stats line**: `🕒 <timing>` then `🧠 <static>[ + <dynamic>]`. Timing is a tick time
-  (µs under 1 ms, ms above), omitted when the module has no measured loop time. Memory is
-  the C++ object size (`classSize`); the `+ <dynamic>` part (`dynamicBytes`, heap) shows only when the
-  module allocated heap. The line shows from expert mode up, and holding it peeks at the rate on every
-  card, the password field's gesture: a tick time inverts to a figure that reads as a frame rate and is
-  not one, so the rate is a press away rather than a mode the cards sit in.
+  (µs under 1 ms, ms above), omitted when the module has no measured loop time. Memory is the C++ object size (`classSize`); the `+ <dynamic>` part (`dynamicBytes`, heap) shows only when the module allocated heap. The line shows from expert mode up, and holding it peeks at the rate on every card, the password field's gesture: a tick time inverts to a figure that reads as a frame rate and is not one, so the rate is a press away rather than a mode the cards sit in.
 - **Reorder is drag-and-drop** (HTML5 DnD) on the whole card, desktop and mobile. A drag starting on
-  an interactive control is canceled in `dragstart` so the control's own gesture wins; a drag is
-  accepted only when source and target share the same `.card-children` container (true siblings).
+  an interactive control is canceled in `dragstart` so the control's own gesture wins; a drag is accepted only when source and target share the same `.card-children` container (true siblings).
 - **Replace (✎) / Delete (×) / drag-reorder** apply to *user-managed children*, a module whose role a
-  container accepts (`acceptsChildRoles`) and that hasn't opted out via `userEditable: false`. Replace
-  swaps for another type at the same position (picker filtered to the module's role; replacement starts
-  at its own defaults). Delete is press-twice confirm. A `userEditable: false` child (e.g.
+  container accepts (`acceptsChildRoles`) and that hasn't opted out via `userEditable: false`. Replace swaps for another type at the same position (picker filtered to the module's role; replacement starts at its own defaults). Delete is press-twice confirm. A `userEditable: false` child (e.g.
   PreviewDriver) shows none of these, it's fixed-shape like a code-wired child.
 - **`+ add child`** renders in a parent's footer when its type declares a non-empty
   `acceptsChildRoles`; it opens the [type picker](#type-picker) filtered to those roles.
 
 ## Control types
 
-Auto-rendered by `controls[].type`, adding a MoonModule with these types needs no UI change. (The
-descriptors themselves, and their storage shape, are [Control](moxygen/Control.md) /
-[coding-standards § store values in their native shape](../../contributing/coding-standards.md); this table is
-how each *renders*.)
+Auto-rendered by `controls[].type`, adding a MoonModule with these types needs no UI change. (The descriptors themselves, and their storage shape, are [Control](moxygen/Control.md) / [coding-standards § store values in their native shape](../../contributing/coding-standards.md); this table is how each *renders*.)
 
 | Type | Element | Interaction | Debounce |
 |---|---|---|---|
@@ -163,40 +125,26 @@ how each *renders*.)
 | `button` | clickable button | sends value = 1 on click | none |
 
 - **`filepath` stores a NAME, not a body.** The value is a ~40-byte reference that travels through
-  `/api/control` like any text control; the file's contents move over `GET`/`POST /api/file`,
-  because that is the only route allowed to exceed the request buffer (everything else returns 413).
-  The module declares where its files live and which to offer (`addFilePath(name, buf, size,
-  pick)`, a {directory, extension, template} triple), so the UI lists a directory, filters it, and
-  seeds a new file without knowing what kind of file it holds. Saving is
-  all it takes: a written file asks the module tree to re-derive, so whatever was built from that
-  file rebuilds itself, with no second request from the browser.
-  Editing is the same code the File Manager's modal editor uses, mounted inline instead of in a
-  `<dialog>`, so both share one set of guards (a binary or truncated file loads read-only rather
-  than risking a lossy re-save).
+  `/api/control` like any text control; the file's contents move over `GET`/`POST /api/file`, because that is the only route allowed to exceed the request buffer (everything else returns 413).
+  The module declares where its files live and which to offer (`addFilePath(name, buf, size, pick)`, a {directory, extension, template} triple), so the UI lists a directory, filters it, and seeds a new file without knowing what kind of file it holds. Saving is all it takes: a written file asks the module tree to re-derive, so whatever was built from that file rebuilds itself, with no second request from the browser.
+  Editing is the same code the File Manager's modal editor uses, mounted inline instead of in a `<dialog>`, so both share one set of guards (a binary or truncated file loads read-only rather than risking a lossy re-save).
 
 - **Reset-to-default (↺)** appears next to controls whose default is known (captured from a fresh
-  probe instance per type, emitted in `/api/types`); dim when value == default, clicking sends the
-  default.
+  probe instance per type, emitted in `/api/types`); dim when value == default, clicking sends the default.
 - **Hidden controls.** `hidden: true` on a descriptor → the UI skips rendering it (conditional
   controls like NetworkModule's static-IP fields under DHCP).
 - **Password controls are obfuscated, not encrypted.** A `password` control serializes in `/api/state`
-  as `{"type":"password","value":"<encoded>"}` where `<encoded>` is the password XOR'd with a fixed key
-  then base64-encoded; the UI decodes it (`decodePassword()`) so the input holds the real value (masked,
-  hold-to-peek). A **first line of defense only**: the XOR key is a shared constant in both the firmware
-  and `app.js`, so it's trivially reversible, it stops the password being plainly readable in a raw
-  `curl /api/state`, not a determined reader.
+  as `{"type":"password","value":"<encoded>"}` where `<encoded>` is the password XOR'd with a fixed key then base64-encoded; the UI decodes it (`decodePassword()`) so the input holds the real value (masked, hold-to-peek). A **first line of defense only**: the XOR key is a shared constant in both the firmware and `app.js`, so it's trivially reversible, it stops the password being plainly readable in a raw `curl /api/state`, not a determined reader.
 
 ## Type picker
 
-One picker serves **add** (`+ add child`) and **replace** (the ✎ button), rendered inline in the card
-(not a modal).
+One picker serves **add** (`+ add child`) and **replace** (the ✎ button), rendered inline in the card (not a modal).
 
 - **Role filter.** Add mode filters to the parent's `acceptsChildRoles` (declared per-type in
   `/api/types`, the UI hardcodes no container→role map); replace mode filters to the target's own role.
 - **Emoji tag chips.** A row of toggle chips above the list, one per distinct emoji across the filtered
   types (sources + meanings: [the tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend)).
-  The UI treats `tags` as opaque, splits it into grapheme clusters, one chip each. Toggling narrows
-  with **AND** logic (a type shows only if it carries every active chip).
+  The UI treats `tags` as opaque, splits it into grapheme clusters, one chip each. Toggling narrows with **AND** logic (a type shows only if it carries every active chip).
 - **Search box**: substring match on type name; combines with chips (both must match).
 - **Keyboard nav**: type to filter, ↓ to enter list, ↑↓ to move, Enter to confirm, Esc to cancel.
 - **Confirm / Cancel** at the bottom (confirm reads `create` or `replace`); double-click a row to
@@ -211,15 +159,11 @@ One picker serves **add** (`+ add child`) and **replace** (the ✎ button), rend
 - **Module nesting** by progressively lighter card backgrounds (depth 0 / 1 / 2).
 - **Responsive breakpoint** at 820 px.
 - **Color semantics** consistent across the app: green (`#22c55e`/`#6ee7b7`) = connected/ok/pass; red
-  (`#f87171`) = error/fail/crashed/delete; purple (`#a78bfa`/`#c4b5fd`) = accent/brand/active/value;
-  gray (`#6b7280`/`#9ca3af`/`#4b5563`) = secondary/muted.
+  (`#f87171`) = error/fail/crashed/delete; purple (`#a78bfa`/`#c4b5fd`) = accent/brand/active/value; gray (`#6b7280`/`#9ca3af`/`#4b5563`) = secondary/muted.
 
 ## Domain preview channel
 
-The UI dedicates a binary slot on its WebSocket, separate from the JSON state updates, for a
-domain-specific preview frame; the engine pushes one per render and the UI hands it to a domain
-renderer. Generic shape: `[type-byte] [domain header] [payload]`, the first byte selects the
-renderer, everything after is the domain's choice; the UI ignores types it doesn't recognise.
+The UI dedicates a binary slot on its WebSocket, separate from the JSON state updates, for a domain-specific preview frame; the engine pushes one per render and the UI hands it to a domain renderer. Generic shape: `[type-byte] [domain header] [payload]`, the first byte selects the renderer, everything after is the domain's choice; the UI ignores types it doesn't recognise.
 
 - The canvas is sticky just below the status bar, scroll-shrinks 0→50 % over 300 px.
 - `width: 100%` + `aspect-ratio: 1 / 1` derives the height from the column width; `max-height: 50vh`
@@ -229,34 +173,26 @@ renderer, everything after is the domain's choice; the UI ignores types it doesn
 - WebGL clear color `(0, 0, 0, 0)`, a transparent canvas blends into either theme with no per-frame
   color work.
 
-The light-domain renderer (WebGL point cloud, the `0x03`/`0x02` frame format, orbit camera,
-downsampling) is [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui) /
-[PreviewDriver](../light/moxygen/PreviewDriver.md).
+The light-domain renderer (WebGL point cloud, the `0x03`/`0x02` frame format, orbit camera, downsampling) is [MoonCore, the web UI](../../explanation/architecture/mooncore.md#web-ui) / [PreviewDriver](../light/moxygen/PreviewDriver.md).
 
 ## State updates, the no-rebuild contract
 
-When a WS state snapshot arrives, for each module in the payload:
+When a WS state message arrives, the full tree on connect or a patch after it, for each module in the payload:
 - If `controls.deviceName` exists, update the status-bar device name and `document.title`.
 - For each `controls[key]`: look up the matching `<input>`/`<span>`/`<progress>`/`<select>` by
-  `[data-mid][data-key]`; **skip** if `Date.now() - dragTs[mid:key] < 1000` (user interacting);
-  otherwise patch the value in place (sliders update input + `value-display`; toggles set `checked`;
-  selects set `value`). Then update the reset-to-default "at default?" state and the timing display.
+  `[data-mid][data-key]`; **skip** if `Date.now() - dragTs[mid:key] < 1000` (user interacting); otherwise patch the value in place (sliders update input + `value-display`; toggles set `checked`; selects set `value`). Then update the reset-to-default "at default?" state and the timing display.
 
-The DOM is **never rebuilt** during a state update. A full re-render happens only on (a) initial load
-and (b) a mutation that changes the tree shape (`/api/control` for a Select that triggers
-`rebuildControls`; `/api/modules` add/delete/move/replace).
+The DOM is **never rebuilt** during a state update. A full re-render happens only on (a) initial load and (b) a mutation that changes the tree shape (`/api/control` for a Select that triggers `rebuildControls`; `/api/modules` add/delete/move/replace).
 
 **Object identity across pushes.** Every WS push replaces `state` with a fresh JSON tree, so a
-card-render closure holding a `mod` reference goes stale within ~1 s. Any "which index am I now?"
-lookup must use `findIndex(c => c.name === mod.name)`, not `indexOf(mod)`.
+card-render closure holding a `mod` reference goes stale within ~1 s. Any "which index am I now?" lookup must use `findIndex(c => c.name === mod.name)`, not `indexOf(mod)`.
 
 ## Communication client behavior
 
-The endpoints, the `/ws` frame shape, and the streaming state sink are owned by
-[HttpServerModule](moxygen/HttpServerModule.md); `/api/control` is `{module, control, value}`. The
+The endpoints, the `/ws` frame shape, and the streaming state sink are owned by [HttpServerModule](moxygen/HttpServerModule.md); `/api/control` is `{module, control, value}`. The
 **browser-side** socket behavior the server doesn't dictate:
 
-- URL `ws://<host>/ws` (same port as HTTP); the server pushes a full-state snapshot ~1/s.
+- URL `ws://<host>/ws` (same port as HTTP); the server sends the full state on connect, then a patch of what changed each second.
 - Client sends `"ping"` every 25 s as keepalive (Safari kills idle sockets otherwise).
 - Auto-reconnect on close with exponential backoff (500 ms → 5 s ceiling).
 - Pause on `document.visibilityState === 'hidden'`, resume on `pageshow` (Safari bfcache survival).
@@ -272,8 +208,4 @@ No other client state persists; reorder, control values, etc. all live on the de
 
 ## Source
 
-Hand-maintained files: [index.html](../../../../src/ui/index.html) ·
-[app.js](../../../../src/ui/app.js) · [style.css](../../../../src/ui/style.css) ·
-[preview3d.js](../../../../src/ui/preview3d.js) ·
-[install-picker.js](../../../../src/ui/install-picker.js). Embedded into firmware at build time via
-[embed_ui.cmake](../../../../src/ui/embed_ui.cmake).
+Hand-maintained files: [index.html](../../../../src/ui/index.html) · [app.js](../../../../src/ui/app.js) · [style.css](../../../../src/ui/style.css) · [preview3d.js](../../../../src/ui/preview3d.js) · [install-picker.js](../../../../src/ui/install-picker.js). Embedded into firmware at build time via [embed_ui.cmake](../../../../src/ui/embed_ui.cmake).
diff --git a/docs/moonmodules/light/drivers.md b/docs/moonmodules/light/drivers.md
index da586f09..40a95d22 100644
--- a/docs/moonmodules/light/drivers.md
+++ b/docs/moonmodules/light/drivers.md
@@ -10,13 +10,14 @@ Several drivers can share one buffer, each driving its own slice. Every driver s
 
 ### Shared 💫 · every driver
 
-Added once by [`DriverBase`](moxygen/DriverBase.md) so no driver re-implements it: a per-driver **output correction** (how this driver's slice looks) and a **source window** (which slice of the shared buffer it reads). Every driver card leads with this block; its own controls follow.
+The block every driver card opens with, shown here on RMT LED. Added once by [`DriverBase`](moxygen/DriverBase.md) so no driver re-implements it: a per-driver **output correction** (how this driver's slice looks) and a **source window** (which slice of the shared buffer it reads). A driver card leads with whichever half applies, and its own controls follow. Hue and Preview correct nothing, since the fixture and the browser do it; HUB75, Preview, NDI and HLS take the whole buffer rather than a window.
 
-<img src="../../assets/light/drivers/RmtLedDriver.png" width="300" alt="Shared driver controls: localBrightness, lightPreset, whiteMode, start, count">
+<img src="../../assets/light/drivers/RmtLedDriver.png" width="300" alt="The shared block at the top of a driver card, here on RMT LED">
 
 - `localBrightness`: this driver's dim (0–255), multiplied with the global brightness.
+- `curve`: brightness to output: `CIE 1931`, `gamma 2.2`, `gamma 2.8`, or `linear`.
 - `lightPreset`: the [light preset](supporting.md) applied per light, for order and white.
-- `whiteMode`: how W is derived on an RGBW strip, when the preset carries a W channel.
+- `whiteMode`: how W is derived, shown when the preset carries a W channel.
 - `start`: first light of the shared buffer this driver reads (default `0`).
 - `count`: how many lights from `start`. **Blank drives all of them.**
 
@@ -30,24 +31,46 @@ Detail: [technical](moxygen/DriverBase.md)
 <a id="moonled"></a>
 <a id="parlioled"></a>
 
-### LED driver 💫 · wire
+### RMT LED 💫 · wire
 
-Addressable WS2812B-class LEDs over a wire: **RMT** for a few strands, **ParallelLedDriver** for up to 16 clocked out at once. Which peripheral drives the parallel bus, and what each one buys, is in [the details below](#led-driver-details).
+Addressable WS2812B-class LEDs on a handful of strands, clocked by the chip's RMT peripheral. The classic choice, and the one that needs no DMA peripheral: for more than a few strands see [Parallel LED](#parallel-led) below. What the peripherals buy is in [the details below](#parallel-led-details).
 
-<img src="../../assets/light/drivers/RmtLedDriver.png" width="300" alt="LED output driver controls">
+<img src="../../assets/light/drivers/RmtLedDriver.png" width="300" alt="RMT LED driver controls">
+
+Plus the [shared controls](#shared-driver-controls) above:
+
+- `pins`: data GPIO list: `18,17,16`, or ranges like `20-23`, mixed freely. Empty idles until set.
+- `ledsPerPin`: lights per strand. Blank splits evenly, one number to all, a list per strand.
+- `timing`: the wire bit rate. A 12V WS2811 needs `400kHz`; the default suits the rest.
+- `t0hNs` / `t1hNs` / `periodNs`: the bit timing itself, shown when `timing` is `custom`.
+- 🔧 `loopbackTest`: a TX to RX self-test, verdict in the status field.
+- 🔧 `loopbackTxPin` / `loopbackRxPin` / `loopbackFrame`: its pins, and whether a whole frame is sent.
+
+Tests: [RMT](../../reference/tests/unit-tests.md#rmtleddriver)
+
+Detail: [technical](moxygen/RmtLedDriver.md)
+
+<a id="parallel-led"></a>
+
+### Parallel LED 💫 · wire
+
+The same LEDs, up to 16 strands clocked out at once by a DMA peripheral, which is what makes a large wall possible. Which peripheral drives the bus, and what each buys, is in [the details below](#parallel-led-details).
+
+<img src="../../assets/light/drivers/ParallelLedDriver.png" width="300" alt="Parallel LED driver controls">
 
 Plus the [shared controls](#shared-driver-controls) above:
 
 - `pins`: data GPIO list: `18,17,16`, or ranges like `20-23`, mixed freely. Empty idles until set.
 - `ledsPerPin`: lights per strand. Blank splits evenly, one number to all, a list per strand.
-- `timing` (RMT only): the wire bit rate. A 12V WS2811 needs `400kHz`; the default suits the rest.
 - `peripheral`: the DMA peripheral, filtered to what the chip supports. It divides the card.
-- `doubleBuffer`, `pinExpander`, bus pins: shown per peripheral, so the set changes with it.
+- `doubleBuffer`, `pinExpander`, `latchPin`, bus pins: shown per peripheral.
+- read-only `frameTime`: what one frame costs on the wire, the headroom the pipeline has left.
 - 🔧 `loopbackTest`: a TX to RX self-test, verdict in the status field.
+- 🔧 `loopbackStrand` / `loopbackIntrusive`: which strand it tests, and whether it may disturb output.
 
-Tests: [RMT](../../reference/tests/unit-tests.md#rmtleddriver) · [shared + peripherals](../../reference/tests/unit-tests.md#parallelleddriver)
+Tests: [shared + peripherals](../../reference/tests/unit-tests.md#parallelleddriver)
 
-Detail: [RMT](moxygen/RmtLedDriver.md) · [Parallel](moxygen/ParallelLedDriver.md) · peripherals: [i80](moxygen/I80Peripheral.md) · [MoonI80](moxygen/MoonI80Peripheral.md) · [Parlio](moxygen/ParlioPeripheral.md)
+Detail: [technical](moxygen/ParallelLedDriver.md) · peripherals: [i80](moxygen/I80Peripheral.md) · [MoonI80](moxygen/MoonI80Peripheral.md) · [Parlio](moxygen/ParlioPeripheral.md)
 
 <a id="hub75"></a>
 
@@ -117,7 +140,7 @@ Detail: [technical](moxygen/PanelCardDriver.md)
 
 <img src="../../assets/light/drivers/HueDriver.png" width="300" alt="A HueDriver in the UI">
 
-Drives **Philips Hue bulbs as pixels**: each color bulb in the driver's window becomes one pixel, pushed to the bridge over its HTTP API. Paced to the bridge's ~10 cmd/s limit, so smooth ambient color, not strobing.
+Drives **Philips Hue bulbs as pixels**: each color bulb in the driver's window becomes one pixel, pushed to the bridge over its HTTP API. Paced to the bridge's ~10 cmd/s limit, so smooth ambient color, not strobing. Up to 32 bulbs, which is a LAN's worth and bounds the driver's memory.
 
 - `bridgeIp`: the bridge's LAN IPv4.
 - `appKey`: the Hue app key; filled by `pair`, persisted.
@@ -140,7 +163,7 @@ Streams a true-shape 3D preview to the web UI as a **point list**, only the real
 
 It streams on its own WebSocket channel so a large frame never delays the control plane, and only while a viewer is watching. Moving heads show their beam. How the rate and detail trade off is in [the details below](#preview-details).
 
-- `targetFps`: the rate to aim for (default 24, 1–60). **Lower for detail, raise for smoothness.**
+- `targetFps`: the rate to aim for (default 24, 1–25). **Lower for detail, raise for smoothness.**
 
 [Tests](../../reference/tests/unit-tests.md#previewdriver)
 
@@ -174,7 +197,7 @@ The grid becomes the frame, output correction applied. Latency is HLS's own, **2
 - `targetFps`: encode-rate ceiling (default 30, 1–120). Also the bandwidth knob: bitrate is derived.
 - `scale`: video pixels per light (0 = auto). Each light is a solid block, never a blur.
 - `encoder`: which ffmpeg encoder (desktop only).
-- read-only: `url` to play, plus a status line for state, drops, or why the encoder stopped.
+- read-only: `url` to play. The card's status line carries state, drops, or why the encoder stopped.
 
 Detail: [technical](moxygen/HlsDriver.md) · [the transport-stream muxer](moxygen/MpegTs.md)
 
@@ -185,9 +208,9 @@ Detail: [technical](moxygen/HlsDriver.md) · [the transport-stream muxer](moxyge
 
 **`start` and `count` are how several drivers share one buffer.** Blank `count` drives every light; a number drives only that slice. An onboard status LED takes `start 0, count 1` while the main strip runs from `start 1`, both reading the same buffer.
 
-<a id="led-driver-details"></a>
+<a id="parallel-led-details"></a>
 
-## LED driver, details
+## Parallel LED, details
 
 **`doubleBuffer`: leave it on.** The driver encodes the next frame into a second DMA buffer while the current one clocks out, so a tick costs `max(encode, wire)` rather than `encode + wire`. Measured on a P4 at 16x256 lights it lifted the whole board from about 48 to 76 fps, moving a 7.7 ms wire into background DMA. It costs one extra DMA buffer and one frame of output latency, roughly 8 to 20 ms, which sits inside the perceptual audio-to-visual window and is small next to the wire itself. There is no setup, audio-reactive included, that should turn it off for latency. A board whose second buffer will not fit degrades to the synchronous path on its own.
 
diff --git a/docs/moonmodules/light/effects.md b/docs/moonmodules/light/effects.md
index d9f50751..1491f9a8 100644
--- a/docs/moonmodules/light/effects.md
+++ b/docs/moonmodules/light/effects.md
@@ -19,18 +19,21 @@ Effects are built from the shared [power functions](power-functions.md): the dra
 Emitters pouring color into a flow that carries and folds it. There is no velocity field: one noise value shifts each row sideways, one shifts each column, and the two shears compose into a swirling current. A large grid is steered by a few hundred numbers, which is why it runs where a solver does not.
 
 - `speed`: how fast the emitters travel.
-- `flow`: how far a row or column is pushed: the current's strength.
+- `flow`: how far a row or column is pushed: the shear amount.
 - `flowSpeed`: how fast the flow itself drifts and reverses.
-- `scale`: the flow's spatial frequency: a few broad bands or many fine ones.
+- `scale`: the profiles' spatial frequency: a few broad bands or many fine ones.
 - `persistence`: how long color survives, as a half-life.
 - `colorSpeed`: how fast the emitters walk the palette.
 - `size`: the orbit's radius and the Lissajous figure's reach.
 - `mode`: all three emitters, or one at a time to see what each contributes.
+- `flowType`: the field carrying them: `Noise`, `Radial out`, `Radial in`.
 
 Compare with [Fluid](#fluid): the solver when the medium is the subject, this when the color is.
 
 Origin: MoonLight · concept by [Stefan Petrick](https://github.com/StefanPetrick), composition by Jeff (mindful_stone / [4wheeljive](https://github.com/4wheeljive)) in [FlowFields](https://github.com/4wheeljive/FlowFields/blob/main/src/flows/flow_noise.h) · via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Effects/E_FastLED.h)
 
+Detail: [technical](moxygen/ColorTrailsEffect.md)
+
 <a id="distortionwaves"></a>
 
 ### DistortionWaves 💫 · 2D
@@ -115,8 +118,10 @@ Detail: [technical](moxygen/LavaLampEffect.md)
 
 Sweeps axis-aligned planes in sync; red/green/blue name the X/Y/Z axis: a preview-orientation test pattern.
 
-- `speed`: sweep BPM.
-- `axis`: which plane sweeps (`all`, `x (red)`, `y (green)`, `z (blue)`).
+- `mode`: `lines` sweeps a plane, `panel dots` walks a dot per panel.
+- `speed`: sweep BPM, in `lines`.
+- `axis`: which plane sweeps: `all`, `x`, `y` or `z`, in `lines`.
+- `panelW` / `panelH`: the panel the dot walks, in `panel dots`.
 
 Origin: MoonLight · via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Effects/E_MoonLight.h)
 
@@ -162,7 +167,7 @@ Detail: [technical](moxygen/ParticlesEffect.md)
 
 <a id="plasma"></a>
 
-### Plasma 💫🦅 · 2D/3D
+### Plasma 💫🦅 · 3D
 
 <img src="../../assets/light/effects/PlasmaEffect.gif" width="300" alt="Plasma effect preview">
 
@@ -186,6 +191,7 @@ Detail: [technical](moxygen/PlasmaEffect.md)
 
 An algorithmic palette pattern driven by two beat oscillators (a macro and a micro mutator) whose frequencies and ranges reshape the hue field over time.
 
+- `speed`: how fast the pattern advances.
 - `macroMutatorFreq` / `Min` / `Max`: the coarse mutator's beat rate and range.
 - `microMutatorFreq` / `Min` / `Max`: the fine mutator's beat rate and range.
 
@@ -272,7 +278,7 @@ Detail: [technical](moxygen/RipplesEffect.md)
 A 3D Rubik's Cube projected onto the volume: it scrambles, then plays its solution back one turn at a time, the six faces in their standard colors.
 
 - `turnsPerSecond`: how fast the cube turns.
-- `cubeSize`: the cube order (2×2 up to 8×8).
+- `cubeSize`: the cube order (1×1 up to 8×8).
 - `randomTurning`: turn endlessly at random instead of scramble-then-solve.
 - `usePalette`: color the faces from the palette, not the classic colors.
 
@@ -302,10 +308,10 @@ Physics runs on elapsed time, so the same settings behave identically at any fra
 
 Origin: projectMM original, on the WLED Particle System's firework family by Damian Schneider / [@DedeHai](https://github.com/DedeHai)
 
-<a id="fishtank"></a>
-
 Detail: [technical](moxygen/FireworksEffect.md)
 
+<a id="fishtank"></a>
+
 ### Fish Tank 💫🎶✨👾 · 2D
 
 <img src="../../assets/light/effects/FishTankEffect.gif" width="300" alt="Fish Tank effect preview">
@@ -323,6 +329,8 @@ Uses the global palette, each fish's band a paler version of its own body color.
 
 Origin: projectMM original; inspired by the aquarium screensavers of the After Dark era, the pixel art drawn fresh for this effect
 
+Detail: [technical](moxygen/FishTankEffect.md)
+
 <a id="flyingtoasters"></a>
 
 ### Flying Toasters 💫🎶✨👾 · 2D
@@ -341,6 +349,8 @@ The sprites carry their own colors (chrome, wing, crust), so the global palette
 
 Origin: projectMM original; inspired by After Dark's Flying Toasters (Berkeley Systems, 1989), suggested by Frank ([softhack007](https://github.com/softhack007)): the pixel art here is drawn fresh for this effect
 
+Detail: [technical](moxygen/FlyingToastersEffect.md)
+
 <a id="fixedpoint"></a>
 
 ### FixedPoint 💫🖌️ · 2D
@@ -387,17 +397,17 @@ A fixture chain is one-dimensional, so lay the rig out as a **1 x N** grid.
 
 Origin: projectMM original
 
-<a id="pacman"></a>
-
 Detail: [technical](moxygen/MovingHeadEffect.md)
 
+<a id="pacman"></a>
+
 ### Pacman 💫🎶✨👾 · 2D
 
 <img src="../../assets/light/effects/PacmanEffect.gif" width="300" alt="Pacman effect preview">
 
 The arcade cast crossing a light wall: Pacman chomps his way along while the four ghosts drift past, each in its own color, wrapping around the edges forever.
 
-In this first iteration the characters travel independently and do not notice each other. The maze, the pellets and the chase are the next step.
+The characters travel independently, each on its own path.
 
 - `pacmen`: how many Pacmen (0-4).
 - `ghosts`: how many ghosts (0-8); the arcade cast is four.
@@ -409,6 +419,8 @@ Pacman keeps his yellow; the ghosts take their colors from the active palette.
 
 Origin: projectMM original; inspired by Namco's Pac-Man (1980), the pixel art drawn fresh for this effect
 
+Detail: [technical](moxygen/PacmanEffect.md)
+
 <a id="spaceinvaders"></a>
 
 ### Space Invaders 💫🎵👾 · 2D
@@ -427,6 +439,8 @@ The invaders take their body color from the active palette.
 
 Origin: projectMM original, after Taito's Space Invaders (1978)
 
+Detail: [technical](moxygen/SpaceInvadersEffect.md)
+
 <a id="spritefountain"></a>
 
 ### Sprite Fountain 💫🎶✨👾 · 2D
@@ -467,10 +481,10 @@ Uses the global palette.
 
 Origin: projectMM original, after Atari's Pong (1972)
 
-<a id="aurora"></a>
-
 Detail: [technical](moxygen/PongEffect.md)
 
+<a id="aurora"></a>
+
 ### Aurora 💫🖌️🌫️🎡 · 3D
 
 <img src="../../assets/light/effects/AuroraEffect.gif" width="300" alt="Aurora effect preview">
@@ -490,10 +504,10 @@ Several noise fields, each on its own clock, read in polar coordinates and compo
 
 Origin: projectMM original, in the shader vocabulary Stefan Petrick made recognizable in the LED world
 
-<a id="ballpit"></a>
-
 Detail: [technical](moxygen/AuroraEffect.md)
 
+<a id="ballpit"></a>
+
 ### Ballpit 💫✨ · 2D
 
 <img src="../../assets/light/effects/BallpitEffect.gif" width="300" alt="Ballpit effect preview">
@@ -511,10 +525,10 @@ Collisions are the one non-linear part of the particle kernel, so the pool is sm
 
 Origin: projectMM original, on the WLED Particle System's ballpit family by Damian Schneider / [@DedeHai](https://github.com/DedeHai)
 
-<a id="dissolve"></a>
-
 Detail: [technical](moxygen/BallpitEffect.md)
 
+<a id="dissolve"></a>
+
 ### Dissolve 💫 · 2D
 
 <img src="../../assets/light/effects/DissolveEffect.gif" width="300" alt="Dissolve effect preview">
@@ -528,10 +542,10 @@ Two color fields trade places pixel by pixel in an order that looks random but i
 
 Origin: projectMM original, on the classic dissolve transition in its position-addressed (shader) form
 
-<a id="echo"></a>
-
 Detail: [technical](moxygen/DissolveEffect.md)
 
+<a id="echo"></a>
+
 ### Echo 💫✨ · 2D
 
 <img src="../../assets/light/effects/EchoEffect.gif" width="300" alt="Echo effect preview">
@@ -548,10 +562,10 @@ Shows that feedback is not a primitive: once the grid can be read as a texture (
 
 Origin: projectMM original, on video feedback and the standard texture-feedback shader shape
 
-<a id="spectrum"></a>
-
 Detail: [technical](moxygen/EchoEffect.md)
 
+<a id="spectrum"></a>
+
 ### Spectrum 💫🎶 · 2D
 
 <img src="../../assets/light/effects/SpectrumEffect.gif" width="300" alt="Spectrum effect preview">
@@ -568,10 +582,10 @@ The asymmetry is the whole point; a symmetric follower either misses the hit or
 
 Origin: projectMM original, on standard VU/PPM meter ballistics and WLED's GEQ band mapping
 
-<a id="truchet"></a>
-
 Detail: [technical](moxygen/SpectrumEffect.md)
 
+<a id="truchet"></a>
+
 ### Truchet 💫🖌️ · 2D
 
 <img src="../../assets/light/effects/TruchetEffect.gif" width="300" alt="Truchet effect preview">
@@ -589,10 +603,10 @@ A maze of interlocking arcs that never repeats, drawn without storing a single t
 
 Origin: projectMM original, on Sébastien Truchet's 1704 tiling and the standard shader fract/hash/smoothstep idiom
 
-<a id="fluid"></a>
-
 Detail: [technical](moxygen/TruchetEffect.md)
 
+<a id="fluid"></a>
+
 ### Fluid 💫🖌️🌊💨 · 3D
 
 <img src="../../assets/light/effects/FluidEffect.gif" width="300" alt="Fluid effect preview">
@@ -610,10 +624,10 @@ On a cube every depth slice is its own medium, so the slices differ. Sized for t
 
 Origin: projectMM original, after Stam 1999 "Stable Fluids"
 
-<a id="nebula"></a>
-
 Detail: [technical](moxygen/FluidEffect.md)
 
+<a id="nebula"></a>
+
 ### Nebula 💫🖌️💨🌫️ · 3D
 
 <img src="../../assets/light/effects/NebulaEffect.gif" width="300" alt="Nebula effect preview">
@@ -632,10 +646,10 @@ Held at 16 bits and dithered on the way out, which keeps a slow fade smooth.
 
 Origin: projectMM original, composing the noise-field and curl-flow kernels: the contrast window is Aurora's, in the shader vocabulary Stefan Petrick made recognizable in the LED world, and the flow is Bridson's curl noise (SIGGRAPH 2007)
 
-<a id="trails"></a>
-
 Detail: [technical](moxygen/NebulaEffect.md)
 
+<a id="trails"></a>
+
 ### Trails 💫🖌️💨🌫️ · 3D
 
 <img src="../../assets/light/effects/TrailsEffect.gif" width="300" alt="Trails effect preview">
@@ -652,10 +666,10 @@ The trail plane is 16-bit, which is what lets a tail fade smoothly rather than s
 
 Origin: projectMM original, in the flow-field idiom (4wheeljive's FlowFields, from a Stefan Petrick concept), with Stam's backward advection for the transport
 
-<a id="tunnel"></a>
-
 Detail: [technical](moxygen/TrailsEffect.md)
 
+<a id="tunnel"></a>
+
 ### Tunnel 💫🖌️🌫️🎡 · 3D
 
 <img src="../../assets/light/effects/TunnelEffect.gif" width="300" alt="Tunnel effect preview">
@@ -671,10 +685,10 @@ A texture mapped onto the inside of an infinite tube, so the viewer appears to f
 
 Origin: projectMM original, on the standard demoscene tunnel
 
-<a id="vectorballs"></a>
-
 Detail: [technical](moxygen/TunnelEffect.md)
 
+<a id="vectorballs"></a>
+
 ### VectorBalls 💫🖌️ · 2D
 
 <img src="../../assets/light/effects/VectorBallsEffect.gif" width="300" alt="VectorBalls effect preview">
@@ -691,10 +705,10 @@ Without painter's ordering a far ball paints over a near one and the object turn
 
 Origin: projectMM original, on the Amiga-era demoscene vector-ball effect
 
-<a id="waterripple"></a>
-
 Detail: [technical](moxygen/VectorBallsEffect.md)
 
+<a id="waterripple"></a>
+
 ### WaterRipple 💫🧬 · 2D
 
 <img src="../../assets/light/effects/WaterRippleEffect.gif" width="300" alt="WaterRipple effect preview">
@@ -712,10 +726,10 @@ Distinct from [Ripples](#ripples), which draws clean concentric circles; this be
 
 Origin: projectMM original, on Hugo Elias's water surface algorithm
 
-<a id="raymarch"></a>
-
 Detail: [technical](moxygen/WaterRippleEffect.md)
 
+<a id="raymarch"></a>
+
 ### Raymarch 💫🖌️ · 2D
 
 <img src="../../assets/light/effects/RaymarchEffect.gif" width="300" alt="Raymarch effect preview">
@@ -732,6 +746,8 @@ Compiled only where the chip has a hardware FPU. Cost is per pixel, so `steps` t
 
 Origin: projectMM original, on Iñigo Quilez's raymarching and distance-function articles
 
+Detail: [technical](moxygen/RaymarchEffect.md)
+
 <a id="polarnoise"></a>
 
 ### PolarNoise 💫🖌️🌫️🎡 · 3D
@@ -753,10 +769,10 @@ Cost scales with `octaves` and `warp`: at `warp` > 0 and `octaves` 2 it is rough
 
 Origin: projectMM original, after Stefan Petrick's polar/noise vocabulary and Iñigo Quilez's domain warping
 
-<a id="sdfshapes"></a>
-
 Detail: [technical](moxygen/PolarNoiseEffect.md)
 
+<a id="sdfshapes"></a>
+
 ### SdfShapes 💫🖌️ · 2D
 
 <img src="../../assets/light/effects/SdfShapesEffect.gif" width="300" alt="SdfShapes effect preview">
@@ -774,10 +790,10 @@ Measured on an ESP32-S3 at 128×128: 20 fps, 728 cycles/pixel using the true-dis
 
 Origin: projectMM original, after Iñigo Quilez's distance-function catalogue and polynomial smooth-minimum (iquilezles.org)
 
-<a id="solid"></a>
-
 Detail: [technical](moxygen/SdfShapesEffect.md)
 
+<a id="solid"></a>
+
 ### Solid 💫 · 3D
 
 <img src="../../assets/light/effects/SolidEffect.gif" width="300" alt="Solid effect preview">
@@ -1150,7 +1166,7 @@ A gradient-noise field indexed straight into the palette: the plainest way to tu
 - `scale`: spatial frequency: low is broad blobs, high is fine detail.
 - `bpm`: how fast it moves.
 
-Origin: FastLED · inoise field (Mark Kriegsman); the `morph` form from WLED via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Effects/E_WLED.h), which shipped it as a separate Noise2D effect until the two were merged
+Origin: FastLED · inoise field (Mark Kriegsman); the `morph` form from WLED via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Effects/E_WLED.h)
 
 Detail: [technical](moxygen/NoiseEffect.md)
 
@@ -1206,10 +1222,10 @@ Every beat is a stone dropped in water. A real wave simulation, which gives what
 
 Origin: projectMM original, the two-buffer water simulation (Gomez 2000) driven by the onset detector
 
-<a id="vumeters"></a>
-
 Detail: [technical](moxygen/BeatRipplesEffect.md)
 
+<a id="vumeters"></a>
+
 ### VuMeters 💫🎶🖌️ · 3D
 
 <img src="../../assets/light/effects/VuMetersEffect.gif" width="300" alt="VuMeters effect preview">
@@ -1225,10 +1241,10 @@ The bass needles are heavier than the treble ones, so the low end swings and the
 
 Origin: projectMM original, on the VU ballistics of IEC 60268-17
 
-<a id="radialspectrum"></a>
-
 Detail: [technical](moxygen/VuMetersEffect.md)
 
+<a id="radialspectrum"></a>
+
 ### RadialSpectrum 💫🎶🖌️🎡 · 3D
 
 <img src="../../assets/light/effects/RadialSpectrumEffect.gif" width="300" alt="RadialSpectrum effect preview">
@@ -1243,10 +1259,10 @@ The spectrum as ripples. Each band owns a sector around the center, mirrored lef
 
 Origin: projectMM original, the radial spectrogram on `PolarLut` and the onset detector
 
-<a id="demoreel"></a>
-
 Detail: [technical](moxygen/RadialSpectrumEffect.md)
 
+<a id="demoreel"></a>
+
 ### DemoReel 💫 · 3D
 
 <img src="../../assets/light/effects/DemoReelEffect.gif" width="300" alt="DemoReel effect preview">
diff --git a/docs/moonmodules/light/layouts.md b/docs/moonmodules/light/layouts.md
index 1effc709..10dc3697 100644
--- a/docs/moonmodules/light/layouts.md
+++ b/docs/moonmodules/light/layouts.md
@@ -228,7 +228,7 @@ Detail: [technical](moxygen/MoonLiveLayout.md) · [how the count is known](#moon
 
 A dense 3D grid, row-major (x fastest, then y, then z); every position maps to a light.
 
-- `width` / `height` / `depth` — grid extent on each axis in lights (1–512).
+- `width` / `height` / `depth` — lights per axis (to 3840, 2160 and 512).
 - `serpentine` — every other row runs in reverse, matching a snaked strip.
 
 Origin: MoonLight · via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Layouts/L_MoonLight.h)
diff --git a/docs/moonmodules/light/modifiers.md b/docs/moonmodules/light/modifiers.md
index 7480c392..f59a6331 100644
--- a/docs/moonmodules/light/modifiers.md
+++ b/docs/moonmodules/light/modifiers.md
@@ -1,6 +1,6 @@
 # Modifiers
 
-Every modifier, one block each: its preview, what it does, and what each control means — together. A modifier sits between an [effect](effects.md) and the output: it reshapes *where* pixels land (or masks them) without changing the effect's drawing. Modifiers compose — a [Layer](moxygen/Layer.md) folds its whole modifier stack each rebuild; a *dynamic* modifier (one that overrides `modifyLive`) also runs a per-frame pass. See [ModifierBase](moxygen/ModifierBase.md) for the static-vs-dynamic split. Each block's emoji are its `tags()` (see the [tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend)); **Kind** is static (baked into the mapping at rebuild) or dynamic (per-frame remap). Modifiers are grouped into sections, and each block carries that modifier's preview, behavior, and control descriptions together. (For how this page maps to the source/asset folders, see the [folder-structure decision](../../contributing/documentation-standards.md#module-pages).)
+Every modifier, one block each: its preview, what it does, and what each control means — together. A modifier sits between an [effect](effects.md) and the output: it reshapes *where* pixels land (or masks them) without changing the effect's drawing. Modifiers compose — a [Layer](moxygen/Layer.md) folds its whole modifier stack each rebuild; a *dynamic* modifier (one that overrides `modifyLive`) also runs a per-frame pass, and an *animated* one keeps a static fold but rebuilds it on its own clock. See [ModifierBase](moxygen/ModifierBase.md) for the static-vs-dynamic split. Each block's emoji are its `tags()` (see the [tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend)); **Kind** is static (baked into the mapping at rebuild) or dynamic (per-frame remap). Modifiers are grouped into sections, and each block carries that modifier's preview, behavior, and control descriptions together. (For how this page maps to the source/asset folders, see the [folder-structure decision](../../contributing/documentation-standards.md#module-pages).)
 
 A modifier folds coordinates rather than drawing, so it reaches for very little of the shared [power function](power-functions.md) toolbox — that page states the split and lists which modifiers use what.
 
@@ -28,7 +28,7 @@ Detail: [technical](moxygen/BlockModifier.md)
 
 Masks the layer in a checkerboard: "off" squares are dropped, "on" squares pass through unchanged.
 
-- `size` — checker square edge in lights (≥1).
+- `size` — checker square edge in lights (1–64).
 - `invert` — flip which squares pass through vs are masked.
 
 Origin: MoonLight · by WildCats08 / [@Brandon502](https://github.com/Brandon502) · via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Modifiers/M_MoonLight.h)
@@ -142,7 +142,7 @@ Detail: [technical](moxygen/TransposeModifier.md)
 
 <a id="moonlive"></a>
 
-### MoonLive · dynamic
+### MoonLive · static
 
 <img src="../../assets/light/modifiers/MoonLiveModifier.gif" width="300" alt="MoonLive scripted modifier preview">
 
@@ -159,7 +159,7 @@ Detail: [technical](moxygen/MoonLiveModifier.md) · [what a script transforms](#
 
 <a id="randommap"></a>
 
-### RandomMap · dynamic
+### RandomMap 💫 · animated
 
 <img src="../../assets/light/modifiers/RandomMapModifier.gif" width="300" alt="RandomMap modifier preview">
 
@@ -175,7 +175,7 @@ Detail: [technical](moxygen/RandomMapModifier.md)
 
 <a id="region"></a>
 
-### Region · static
+### Region 💫 · static
 
 <img src="../../assets/light/modifiers/RegionModifier.gif" width="300" alt="Region modifier preview">
 
@@ -191,7 +191,7 @@ Detail: [technical](moxygen/RegionModifier.md)
 
 <a id="rotate"></a>
 
-### Rotate · dynamic
+### Rotate 💫 · dynamic
 
 <img src="../../assets/light/modifiers/RotateModifier.gif" width="300" alt="Rotate modifier preview">
 
diff --git a/docs/moonmodules/light/moonlive.md b/docs/moonmodules/light/moonlive.md
index c90278b0..7c1deeb3 100644
--- a/docs/moonmodules/light/moonlive.md
+++ b/docs/moonmodules/light/moonlive.md
@@ -6,9 +6,7 @@ The library that ships with projectMM is [moonlive/](https://github.com/MoonModu
 
 ## A script is a class
 
-Each script declares a **class**, and the host calls its functions: `tick()` for an effect,
-`placeLights()` for a layout, `modifyLogical()` for a modifier. A function is called when it is
-present and its moment arrives, so which entry points a class defines is what decides what it does.
+Each script declares a **class**, and the host calls its functions: `tick()` for an effect, `placeLights()` for a layout, `modifyLogical()` for a modifier. A function is called when it is present and its moment arrives, so which entry points a class defines is what decides what it does.
 The class name is independent of the file name, the way a C file and the functions in it are.
 
 A class may also define functions of its own and **call them**, including calling itself:
@@ -24,11 +22,8 @@ class CrosshairEffect {
 }
 ```
 
-These are real calls, not pasted-in text: the callee gets its own frame when it runs, which is what
-lets one helper call another and lets a function recurse. Arguments are passed BY VALUE, so a
-function that writes a parameter changes its own copy and the caller's variable is untouched.
-`effects/crosshair.mle` is the worked example; `layouts/sixteen-rings.mll` is the one where the
-arguments earn their place, calling one helper sixteen times with different coordinates.
+These are real calls, not pasted-in text: the callee gets its own frame when it runs, which is what lets one helper call another and lets a function recurse. Arguments are passed BY VALUE, so a function that writes a parameter changes its own copy and the caller's variable is untouched.
+`effects/crosshair.mle` is the worked example; `layouts/sixteen-rings.mll` is the one where the arguments earn their place, calling one helper sixteen times with different coordinates.
 
 ### A script is C++
 
@@ -36,8 +31,7 @@ Every shipped script compiles under `c++ -std=c++20 -fsyntax-only`, pinned by `t
 The language therefore stays a subset rather than drifting into a dialect one feature at a time.
 An editor highlights a script correctly, and a reader brings their C++ intuition to it.
 
-One shape difference is deliberate, and the test bridges exactly that one: a class here needs no
-`public:` and no trailing semicolon.
+One shape difference is deliberate, and the test bridges exactly that one: a class here needs no `public:` and no trailing semicolon.
 
 Anything else a compiler rejects is a divergence, and the test is where it surfaces.
 
@@ -46,8 +40,7 @@ Anything else a compiler rejects is a divergence, and the test is where it surfa
 A script does this the way the compiled module it stands in for does.
 `void` acts and answers nothing, `int` hands back a number, and `string` hands back one of the script's own literals.
 
-`return` leaves a function, with a value or without one. Inside `tick()` a bare `return;` is an
-early exit, which is what a guard wants:
+`return` leaves a function, with a value or without one. Inside `tick()` a bare `return;` is an early exit, which is what a guard wants:
 
 ```
 void tick() {
@@ -56,30 +49,19 @@ void tick() {
 }
 ```
 
-`string` names what comes back rather than introducing a string type: a script returns a literal,
-and building, joining or comparing text is out of scope.
+`string` names what comes back rather than introducing a string type: a script returns a literal, and building, joining or comparing text is out of scope.
 
 ### Members, and which become controls
 
-`byte bpm = 30;` is
-state the script owns: visible in every function, surviving every tick. Naming it in
-`defineControls()` with `addControl("bpm", bpm, 1, 240)` also puts it on the UI as a slider, which is
-the same call a compiled module makes. A member no `addControl` names stays private to the script,
-which is how a stateful effect holds a value the user should not see.
+`byte bpm = 30;` is state the script owns: visible in every function, surviving every tick. Naming it in `defineControls()` with `addControl("bpm", bpm, 1, 240)` also puts it on the UI as a slider, which is the same call a compiled module makes. A member no `addControl` names stays private to the script, which is how a stateful effect holds a value the user should not see.
 
-The default comes from the declaration, the range from the call, and the quoted name is the UI
-label, free to differ from the member's name.
+The default comes from the declaration, the range from the call, and the quoted name is the UI label, free to differ from the member's name.
 
 ### A member can be written
 
-That is what makes it state. `level = level + 10;` assigns, and the
-value is still there on the next tick, because a member lives in storage that outlives the call. A
-loop variable can be assigned too. A [system variable](#system-variables)
-(`width`, `t`, `xPos`) cannot: the engine rewrites it before every call, so the store would vanish.
+That is what makes it state. `level = level + 10;` assigns, and the value is still there on the next tick, because a member lives in storage that outlives the call. A loop variable can be assigned too. A [system variable](#system-variables) (`width`, `t`, `xPos`) cannot: the engine rewrites it before every call, so the store would vanish.
 
-A control CAN be assigned, and the effect is visible rather than surprising: the value moves under
-the slider until the user drags it again. Whether a member is a control is decided by
-`defineControls()` at run time, so the language does not distinguish the two here.
+A control CAN be assigned, and the effect is visible rather than surprising: the value moves under the slider until the user drags it again. Whether a member is a control is decided by `defineControls()` at run time, so the language does not distinguish the two here.
 
 ### Branching
 
@@ -103,37 +85,25 @@ byte  heat[16];         // sixteen elements, all zero to begin with
 
 ### Every variable is declared
 
-That includes a loop's counter. A member states its type, an assignment
-to a name that was never declared is refused, and a `for` writes `for (int i = 0; ...)`. One rule
-with no exception, and the same line C++ would take.
+That includes a loop's counter. A member states its type, an assignment to a name that was never declared is refused, and a `for` writes `for (int i = 0; ...)`. One rule with no exception, and the same line C++ would take.
 
 An index is an arbitrary expression such as `heat[i * 2 + 1]`, and one outside the array is **clamped to the last element**.
 A script computes indices from live control values, so out of range is an ordinary run-time state.
 The fixture shows a repeated last light rather than crashing.
 
-All of a class's members share a small fixed budget (`kCtrlBytes`), so a class that declares more
-than fits is a compile error naming the arena, not a failed allocation while a fixture runs.
+All of a class's members share a small fixed budget (`kCtrlBytes`), so a class that declares more than fits is a compile error naming the arena, not a failed allocation while a fixture runs.
 
-`effects/ember.mle` is the worked example: a heat array that decays and re-ignites, so what it
-draws this frame depends on the last one. That is the line between an effect that evaluates a
-formula and one that runs a simulation, and it is the reason arrays exist. `plasma.mle` would look
-identical if every frame started from scratch; `ember.mle` would go dark.
+`effects/ember.mle` is the worked example: a heat array that decays and re-ignites, so what it draws this frame depends on the last one. That is the line between an effect that evaluates a formula and one that runs a simulation, and it is the reason arrays exist. `plasma.mle` would look identical if every frame started from scratch; `ember.mle` would go dark.
 
 ### Helpers and recursion
 
-Declare a helper above the function that calls it. Only functions already parsed are visible, so
-a call to one declared further down reports `unknown function`. A function can always call itself.
+Declare a helper above the function that calls it. Only functions already parsed are visible, so a call to one declared further down reports `unknown function`. A function can always call itself.
 
-Recursion is bounded. About 30 calls deep, a further call does nothing and returns. A render
-task has a fixed stack, so the alternative to a limit is a device that resets mid-frame. What you
-see if you hit it is the picture being wrong where the recursion stopped, on a device that keeps
-running. Nothing is reported; the exact depth is `kMaxCallDepth`.
+Recursion is bounded. About 30 calls deep, a further call does nothing and returns. A render task has a fixed stack, so the alternative to a limit is a device that resets mid-frame. What you see if you hit it is the picture being wrong where the recursion stopped, on a device that keeps running. Nothing is reported; the exact depth is `kMaxCallDepth`.
 
 ### A script says what it is
 
-`dimensions()` and `tags()` are both optional, both named after the
-member functions a compiled module declares (`Dim dimensions() const override`,
-`const char* tags() const override`), and both read once when the script compiles.
+`dimensions()` and `tags()` are both optional, both named after the member functions a compiled module declares (`Dim dimensions() const override`, `const char* tags() const override`), and both read once when the script compiles.
 
 ```
 class RainEffect {
@@ -144,20 +114,11 @@ class RainEffect {
 }
 ```
 
-`dimensions()` returns 1, 2 or 3, and it decides how the layer EXTRUDES the script. A script that
-returns 1 paints the x=0 column and the framework fans it across the width; one that returns 2
-paints the z=0 slice and the framework copies it through the depth. So a script fills a rig it never
-indexed, and a wrong answer is visible: declare 1 and paint a picture, and only the first column
-survives. A script that stays silent is treated as 2, which is what every script rendered as before
-this existed.
+`dimensions()` returns 1, 2 or 3, and it decides how the layer EXTRUDES the script. A script that returns 1 paints the x=0 column and the framework fans it across the width; one that returns 2 paints the z=0 slice and the framework copies it through the depth. So a script fills a rig it never indexed, and a wrong answer is visible: declare 1 and paint a picture, and only the first column survives. A script that stays silent is treated as 2, which is what every script rendered as before this existed.
 
-`tags()` returns the emoji shown beside the script, so a row in the picker reads like a compiled
-effect's. The vocabulary is shared with the compiled modules: 📊 audio-reactive, ✨ particles,
-🎯 aims moving heads. A script that declares none shows 📝, the mark of a scripted effect.
+`tags()` returns the emoji shown beside the script, so a row in the picker reads like a compiled effect's. The vocabulary is shared with the compiled modules: 📊 audio-reactive, ✨ particles, 🎯 aims moving heads. A script that declares none shows 📝, the mark of a scripted effect.
 
-Both reach the picker before a factory script is downloaded, because the build extracts them from
-the source into the catalog. That copy is for display only: once a script is on the device, the
-compiled script is what decides.
+Both reach the picker before a factory script is downloaded, because the build extracts them from the source into the catalog. That copy is for display only: once a script is on the device, the compiled script is what decides.
 
 ## The five types
 
@@ -300,9 +261,7 @@ Its time axis must be monotonic, so feeding it a `beat()` sawtooth reads as a hi
 `addLight(print(xx), yy, 0)` places the same light and reports what `xx` was.
 
 **Take it out again when the script works.** A serial write blocks, and a script runs on the render
-tick, so a print costs frame time every frame it survives. Each compile grants a short burst and then
-goes quiet, which bounds the damage and gives every edit a fresh window; it does not make a print
-free. No script in this folder ships with one.
+tick, so a print costs frame time every frame it survives. Each compile grants a short burst and then goes quiet, which bounds the damage and gives every edit a fresh window; it does not make a print free. No script in this folder ships with one.
 
 ## Source
 
diff --git a/docs/moonmodules/light/power-functions.md b/docs/moonmodules/light/power-functions.md
index 9acd502b..7e29990c 100644
--- a/docs/moonmodules/light/power-functions.md
+++ b/docs/moonmodules/light/power-functions.md
@@ -1,49 +1,38 @@
 # Power functions
 
-The shared toolbox the light domain is built from: a small set of named, integer-only routines that
-[effects](effects.md) compose into a look. One home per idea — an effect that needs a distance, a
-bar, a noise field or a smooth follower calls the same one every other effect calls, so behavior is
-consistent, the cost is measured once, and a fix reaches everything at once.
+The shared toolbox the light domain is built from: a small set of named, integer-only routines that [effects](effects.md) compose into a look. One home per idea, an effect that needs a distance, a bar, a noise field or a smooth follower calls the same one every other effect calls, so behavior is consistent, the cost is measured once, and a fix reaches everything at once.
 
 Three consumers share this vocabulary, and each uses a different slice of it:
 
-- **[Effects](effects.md)** are the main consumer — they draw, so they reach for nearly all of it.
+- **[Effects](effects.md)** are the main consumer, they draw, so they reach for nearly all of it.
 - **[Modifiers](modifiers.md)** fold coordinates through `modifyLogical` and never draw, so they
-  reach for almost none of it. That asymmetry is the architecture, not a gap: **an effect decides
-  what a pixel looks like, a modifier decides where a pixel comes from.**
+  reach for almost none of it. That asymmetry is the architecture, not a gap: **an effect decides what a pixel looks like, a modifier decides where a pixel comes from.**
 - **[MoonLive](moonlive.md)** scripts reach the same routines through the builtin table
-  (`core/moonlive/MoonLiveBuiltins.h`), which carries plain scalar arguments — so a script sees the
-  flat form of a function, not the C++ callback form a compiled effect can use.
+  (`core/moonlive/MoonLiveBuiltins.h`), which carries plain scalar arguments, so a script sees the flat form of a function, not the C++ callback form a compiled effect can use.
 
-Sources: [draw.h](moxygen/draw.md) (drawing), `core/math16.h` (16-bit math), `core/noise.h` (fields).
-The caller lists below are generated by reading the call sites, so they record what the code does
-rather than what it intends.
+Sources: [draw.h](moxygen/draw.md) (drawing), `core/util/math16.h` (16-bit math), `core/util/noise.h` (fields).
+The caller lists below are generated by reading the call sites, so they record what the code does rather than what it intends.
 
 Many of the names here are **[FastLED](https://github.com/FastLED/FastLED)'s**, deliberately:
-`scale8`, `sin8`, the gradient-palette model (`CRGBPalette16` / `colorFromPalette`), and the
-`beatsin8` / `inoise8` / `qadd8` family are the vocabulary the LED-effect world already shares, so a
-contributor recognises them on sight. projectMM links no part of FastLED: the implementations are
-ours, integer-only and tuned for this render loop, with FastLED credited as the prior art behind
-the convention here and in each primitive's own notes (`core/math8.h` names Mark Kriegsman's
-lib8tion directly).
+`scale8`, `sin8`, the gradient-palette model (`CRGBPalette16` / `colorFromPalette`), and the `beatsin8` / `inoise8` / `qadd8` family are the vocabulary the LED-effect world already shares, so a contributor recognises them on sight. projectMM links no part of FastLED: the implementations are ours, integer-only and tuned for this render loop, with FastLED credited as the prior art behind the convention here and in each primitive's own notes (`core/util/math8.h` names Mark Kriegsman's lib8tion directly).
 
-## Migrating an effect — two steps, in this order
+## Migrating an effect: two steps, in this order
 
-**Step 1, the port: behave identically.** Bringing an effect over from WLED or MoonLight reproduces the original's visual behavior exactly, because the original is the best available description of what the effect should look like. At this stage a difference is a bug, not a variation — pin it with a golden so any drift is visible. Don't get creative with defaults, oscillator math, color mapping, or geometry, and don't silently drop a parameter that *is* the mechanism (the PaintBrush straight-vs-curved-lines bug was a dropped partial-line `length`; Game of Life was wrong the first time by not porting the real algorithm). Study the source for the algorithm, defaults, and visual result, then write our **own** implementation against `EffectBase` and our primitives — carry the behavior forward, don't trace or copy the structure (see [*Industry standards, our own code*](../../../CLAUDE.md#principles)). Credit the origin as prior art in the block below.
+**Step 1, the port: behave identically.** Bringing an effect over from WLED or MoonLight reproduces the original's visual behavior exactly, because the original is the best available description of what the effect should look like. At this stage a difference is a bug, not a variation, pin it with a golden so any drift is visible. Don't get creative with defaults, oscillator math, color mapping, or geometry, and don't silently drop a parameter that *is* the mechanism (the PaintBrush straight-vs-curved-lines bug was a dropped partial-line `length`; Game of Life was wrong the first time by not porting the real algorithm). Study the source for the algorithm, defaults, and visual result, then write our **own** implementation against `EffectBase` and our primitives, carry the behavior forward, don't trace or copy the structure (see [*Industry standards, our own code*](../../../CLAUDE.md#principles)). Credit the origin as prior art in the block below.
 
-**Step 2, the tuning: change it deliberately.** Once the port is faithful it becomes ours to improve. Adopting a [power function](power-functions.md) often makes an effect look better as a side effect — bouncing balls that collide with each other because the physics is now the shared kernel's, a gradient that stops banding because the maths went 16-bit — and that is a real gain, not a regression. The rule is only that the change is deliberate and visible: say what moved, re-baseline the golden in the same commit, and let the product owner judge it on the panel. What is forbidden is drifting silently.
+**Step 2, the tuning: change it deliberately.** Once the port is faithful it becomes ours to improve. Adopting a [power function](power-functions.md) often makes an effect look better as a side effect, bouncing balls that collide with each other because the physics is now the shared kernel's, a gradient that stops banding because the maths went 16-bit, and that is a real gain, not a regression. The rule is only that the change is deliberate and visible: say what moved, re-baseline the golden in the same commit, and let the product owner judge it on the panel. What is forbidden is drifting silently.
 
 ## Used by nearly everything
 
-Not a category so much as the floor: three things almost every effect touches whatever else it does. If you read only one row of this page, read these — an effect that uses none of them is doing something unusual.
+Not a category so much as the floor: three things almost every effect touches whatever else it does. If you read only one row of this page, read these, an effect that uses none of them is doing something unusual.
 
 <div class="mm-pf" markdown="1">
 
 | Power function | What it does | Effects | Modifiers |
 |---|---|---|---|
 | `colorFromPalette` | Reads a color out of the shared palette by index, so every effect follows the user's palette choice | **39 of 47** | — |
-| `draw::pixel` | Writes one pixel, clipped to the grid | **24 of 47** — the rest reach it through a higher-level primitive | — |
-| `BeatPhase` | A BPM phase accumulator that keeps its numerator in 64 bits and divides late, so animation never freezes on sub-millisecond frames | **13** — Dissolve, DistortionWaves, Echo, LavaLamp, Metaballs, Noise, Plasma, PolarNoise, SdfShapes, Sine, Spiral, Tunnel, Wave | — |
+| `draw::pixel` | Writes one pixel, clipped to the grid | **24 of 47**, the rest reach it through a higher-level primitive | — |
+| `BeatPhase` | A BPM phase accumulator that keeps its numerator in 64 bits and divides late, so animation never freezes on sub-millisecond frames | **13**, Dissolve, DistortionWaves, Echo, LavaLamp, Metaballs, Noise, Plasma, PolarNoise, SdfShapes, Sine, Spiral, Tunnel, Wave | — |
 
 </div>
 
@@ -58,13 +47,13 @@ These act on the grid as a surface rather than on a shape. Between them they cov
 | Power function | What it does | Effects | Modifiers |
 |---|---|---|---|
 | `draw::fill` | Fills every light with one color, leaving channels beyond RGB untouched | AudioSpectrum, Blurz, RubiksCube, Solid, Spectrum, Text | — |
-| `draw::fade` | Fades every channel toward black — the trail primitive | **13, through `Layer::fadeToBlackBy`** — Blurz, BouncingBalls, Fireworks, FixedRectangle, FreqSaws, GEQ, GEQ3D, Lissajous, NoiseMeter, PaintBrush, Random, StarField, StarSky | — |
+| `draw::fade` | Fades every channel toward black, the trail primitive | **13, through `Layer::fadeToBlackBy`**, Blurz, BouncingBalls, Fireworks, FixedRectangle, FreqSaws, GEQ, GEQ3D, Lissajous, NoiseMeter, PaintBrush, Random, StarField, StarSky | — |
 | `draw::blur` | Separable box blur across every axis with extent > 1; one call covers 1D, 2D and 3D | Blurz | — |
 | `draw::get` | Reads one pixel back, black outside the grid | Echo, GameOfLife | — |
 | `draw::blendPixel` | Lerps a pixel toward a color by an amount, rather than replacing it | GameOfLife, Tetrix | — |
 | `draw::addPixel` | Adds light to a pixel, saturating instead of wrapping to black | Blurz | — |
-| `draw::scroll` | Shifts the whole grid along an axis, optionally wrapping — the shift register | FreqMatrix | — |
-| `draw::splat` | Draws a point at a fractional position, splitting its light across neighbouring pixels so motion is smooth on a coarse grid | *(no caller yet — the particle kernel is its consumer)* | — |
+| `draw::scroll` | Shifts the whole grid along an axis, optionally wrapping, the shift register | FreqMatrix | — |
+| `draw::splat` | Draws a point at a fractional position, splitting its light across neighbouring pixels so motion is smooth on a coarse grid | *(no caller yet, the particle kernel is its consumer)* | — |
 
 </div>
 
@@ -96,7 +85,7 @@ So an effect whose trail must survive at any framerate keeps its plane wider tha
 
 **Drawing a shape by walking the pixels it covers.**
 
-The classical rasteriser: given endpoints, a center and a radius, or a run length, light exactly the cells the shape passes through. Integer-only and exact, with no distance computed anywhere — which makes these the cheap way to draw when the shape sits on the grid and does not need to move smoothly between pixels.
+The classical rasteriser: given endpoints, a center and a radius, or a run length, light exactly the cells the shape passes through. Integer-only and exact, with no distance computed anywhere, which makes these the cheap way to draw when the shape sits on the grid and does not need to move smoothly between pixels.
 
 Contrast with signed distance fields below: same shapes, opposite approach, different trade-off.
 
@@ -105,10 +94,10 @@ Contrast with signed distance fields below: same shapes, opposite approach, diff
 | Power function | What it does | Effects | Modifiers |
 |---|---|---|---|
 | `draw::line` | A straight line between two points (3D Bresenham) | GEQ3D, PaintBrush | — |
-| `draw::bar` | A run of cells growing from an origin along one axis, colored per cell — the audio-meter staple | AudioSpectrum, GEQ, Spectrum | — |
+| `draw::bar` | A run of cells growing from an origin along one axis, colored per cell, the audio-meter staple | AudioSpectrum, GEQ, Spectrum | — |
 | `draw::fillCircle` | A filled disc (midpoint algorithm), colored per row | Echo | — |
 | `draw::text` | Draws a string, returning its pixel width (`glyph` is reached through it, not called directly) | DemoReel, Text | — |
-| `draw::sprite` | Blits one frame of a small palette-indexed bitmap (index 0 = transparent key), clipped at every edge — the multi-color sibling of `glyph`; movement belongs to the particle pool, never to the sprite | FlyingToasters | — |
+| `draw::sprite` | Blits one frame of a small palette-indexed bitmap (index 0 = transparent key), clipped at every edge, the multi-color sibling of `glyph`; movement belongs to the particle pool, never to the sprite | FlyingToasters | — |
 | `draw::circle` | A circle outline on integer coordinates, exact and symmetric | *(no caller yet)* | — |
 | `draw::rect`, `fillRect` | An axis-aligned rectangle, outlined or filled | *(no caller yet)* | — |
 | `draw::lineAA` | An anti-aliased line (Wu 1991) that splits its light between the two cells straddling the true path | *(no caller yet)* | — |
@@ -119,7 +108,7 @@ Contrast with signed distance fields below: same shapes, opposite approach, diff
 
 **Describing a shape as "how far away is it", then reading a picture out of that number.**
 
-Instead of drawing a circle, an SDF answers *how far is this pixel from the circle's edge* — negative inside, zero on it, positive outside. That one number does far more work than a rasteriser's yes/no: the sign fills the shape, its magnitude gives an anti-aliased edge for free, taking the absolute value turns it into an outline, and two distances combine into a third shape with a single `min` or `smin`.
+Instead of drawing a circle, an SDF answers *how far is this pixel from the circle's edge*, negative inside, zero on it, positive outside. That one number does far more work than a rasteriser's yes/no: the sign fills the shape, its magnitude gives an anti-aliased edge for free, taking the absolute value turns it into an outline, and two distances combine into a third shape with a single `min` or `smin`.
 
 This is what makes shapes composable and smooth-moving. It costs a distance per pixel, so it is the right tool when a shape moves sub-pixel or merges with another, and the wrong one for a static bar.
 
@@ -129,7 +118,7 @@ This is what makes shapes composable and smooth-moving. It costs a distance per
 |---|---|---|---|
 | `sdCircle` | True signed distance to a circle's edge: negative inside, zero on the rim, positive outside | SdfShapes | — |
 | `sdBox` | Signed distance to an axis-aligned box | SdfShapes | — |
-| `smin` | Smooth minimum of two distances — the operator that makes shapes flow together instead of merely overlapping | SdfShapes | — |
+| `smin` | Smooth minimum of two distances, the operator that makes shapes flow together instead of merely overlapping | SdfShapes | — |
 | `coverage` | Turns a distance into 0..255 coverage, which is anti-aliasing for free | SdfShapes | — |
 | `sdCircleSq` | The squared form: same sign contract without the square root, for a plain fill | *(no caller yet)* | — |
 | `sdSegment` | Signed distance to a thick line segment (a capsule) | *(no caller yet)* | — |
@@ -140,7 +129,7 @@ This is what makes shapes composable and smooth-moving. It costs a distance per
 
 **Smooth pseudo-random values across space: everything organic.**
 
-Noise is the source of anything that should look natural rather than drawn — clouds, fire, smoke, water, marbling, drifting color. The defining property is that nearby points get similar values (unlike a raw hash), so the result flows instead of flickering.
+Noise is the source of anything that should look natural rather than drawn, clouds, fire, smoke, water, marbling, drifting color. The defining property is that nearby points get similar values (unlike a raw hash), so the result flows instead of flickering.
 
 One sample is a soft blur; the character comes from composing them. Summing octaves adds structure at every scale, folding the field creases it into flame, and displacing the sample coordinate by another field is what produces the flowing, liquid look.
 
@@ -150,9 +139,9 @@ One sample is a soft blur; the character comes from composing them. Summing octa
 |---|---|---|---|
 | `inoise8` | Perlin gradient noise in 1D, 2D or 3D: a smooth, deterministic pseudo-random field | Noise, NoiseMeter, Wave | `noise` |
 | `fbm8` | Sums noise octaves at doubling frequency and halving amplitude, turning a blur into cloud and terrain structure. Re-widened per octave count, so the field keeps its full range however many are summed | PolarNoise, Tunnel, Aurora, Nebula | `fbm` |
-| `warp8` | Samples noise at a coordinate that noise itself displaced — the flowing, marbled look | PolarNoise, Aurora | `warp` |
+| `warp8` | Samples noise at a coordinate that noise itself displaced, the flowing, marbled look | PolarNoise, Aurora | `warp` |
 | `turbulence8` | Sums the folded absolute value of noise, whose creases read as billowing smoke and flame | *(no caller yet)* | — |
-| `blobCentres` + `blobField` | Orbits N sources on sine paths and sums their inverse-square falloff — the metaball field behind anything fluid or molten | LavaLamp, Metaballs | — |
+| `blobCentres` + `blobField` | Orbits N sources on sine paths and sums their inverse-square falloff, the metaball field behind anything fluid or molten | LavaLamp, Metaballs | — |
 | `curl16` | The perpendicular gradient of a noise potential: a flow field that is divergence-free by construction, so what it carries neither piles up nor drains away | Nebula, Trails | — |
 | `Fluid` | A stable-fluid solver (Stam 1999) in Q16.16: the medium works out its own motion rather than reading it from a function, one independent medium per depth slice | Fluid | — |
 
@@ -179,15 +168,9 @@ The cost is per cell per iteration across several passes, so this is a desktop a
 
 ## Transport
 
-A field says where things go; these carry light along one, frame after frame. The state they move is
-the effect's own plane rather than the layer buffer, held at 16 bits because a value multiplied by
-slightly less than one many times a second has nowhere to go at 8.
+A field says where things go; these carry light along one, frame after frame. The state they move is the effect's own plane rather than the layer buffer, held at 16 bits because a value multiplied by slightly less than one many times a second has nowhere to go at 8.
 
-Each takes a depth, so a cube is the same call as a panel and a panel pays nothing for it (`d = 1`
-reduces to the 2D loop exactly). What that buys on a cube differs by kernel: the noise fields sample
-a genuine third axis, so slices differ rather than one plane repeating, while the transport kernels
-carry light WITHIN each slice and not yet between them. Moving light through a volume needs a
-trilinear sampler and a third velocity component, which is open work.
+Each takes a depth, so a cube is the same call as a panel and a panel pays nothing for it (`d = 1` reduces to the 2D loop exactly). What that buys on a cube differs by kernel: the noise fields sample a genuine third axis, so slices differ rather than one plane repeating, while the transport kernels carry light WITHIN each slice and not yet between them. Moving light through a volume needs a trilinear sampler and a third velocity component, which is open work.
 
 <div class="mm-pf" markdown="1">
 
@@ -204,7 +187,7 @@ trilinear sampler and a third velocity component, which is open work.
 
 **Addressing the grid by angle and radius instead of by x and y.**
 
-Swapping coordinate systems is the cheapest way to change what an effect looks like. Anything radial — rings, spirals, rotation, kaleidoscopes, tunnels, radial wipes, a spectrum bent around a circle — is an ordinary pattern read through polar coordinates rather than a special algorithm.
+Swapping coordinate systems is the cheapest way to change what an effect looks like. Anything radial, rings, spirals, rotation, kaleidoscopes, tunnels, radial wipes, a spectrum bent around a circle, is an ordinary pattern read through polar coordinates rather than a special algorithm.
 
 The 16-bit forms matter here: the 8-bit versions step visibly on a large fixture and their distance is an octagon rather than a circle.
 
@@ -214,10 +197,10 @@ The 16-bit forms matter here: the 8-bit versions step visibly on a large fixture
 |---|---|---|---|
 | `PolarLut` | Every pixel's angle and radius, built once per geometry and read as a table. 34% of a PolarNoise frame on an ESP32-S3, at 2 bytes per pixel (4 at full precision); declines the table when free heap minus the reserve cannot hold it, and the caller computes the address instead | PolarNoise, Spiral, Tunnel, Aurora | — |
 | `atan16` | The angle of a point as a 16-bit turn, smooth enough that a sweep shows no steps on a large fixture | Rings, and the fallback path of every `PolarLut` caller | — |
-| `dist16` | True Euclidean radius — not the octagon `dist8` approximates, and it does not saturate at 255 | Rings, and the fallback path of every `PolarLut` caller | — |
+| `dist16` | True Euclidean radius, not the octagon `dist8` approximates, and it does not saturate at 255 | Rings, and the fallback path of every `PolarLut` caller | — |
 | `kaleido` | Folds an angle into n mirrored wedges, giving any field n-fold symmetry for one modulo | PolarNoise, Tunnel, Aurora | — |
 | `isqrt` | Integer square root with no divide and no float | PaintBrush, WaterRipple | — |
-| `sin8` / `cos8` | The 8-bit oscillators — the internal fast path where a mod-256 result is exactly right | *(the 16-bit forms are the contract)* | **Rotate** |
+| `sin8` / `cos8` | The 8-bit oscillators, the internal fast path where a mod-256 result is exactly right | *(the 16-bit forms are the contract)* | **Rotate** |
 
 </div>
 
@@ -225,7 +208,7 @@ The 16-bit forms matter here: the 8-bit versions step visibly on a large fixture
 
 **How a value changes between frames, and how to get randomness that behaves.**
 
-Two related problems. First, motion: raw linear movement reads as mechanical, so easings shape it, followers smooth it, and peak-hold gives a meter its characteristic instant-rise slow-fall. Second, randomness that is *reproducible* — addressed by position rather than drawn from a stream, so the same pixel gets the same value on every device and every frame.
+Two related problems. First, motion: raw linear movement reads as mechanical, so easings shape it, followers smooth it, and peak-hold gives a meter its characteristic instant-rise slow-fall. Second, randomness that is *reproducible*, addressed by position rather than drawn from a stream, so the same pixel gets the same value on every device and every frame.
 
 The framerate rule lives here too: everything in this group is driven by elapsed time, never by frame count ([architecture](../../explanation/architecture/moonlight.md#effects)). The shapes themselves live in [oscillators](../core/moxygen/oscillators.md), which traces one over a cycle.
 
@@ -237,7 +220,7 @@ The framerate rule lives here too: everything in this group is driven by elapsed
 | `OscillatorBank` | N independent low-frequency oscillators advanced once per frame and read per pixel, each with its own rate, shape, range and phase offset. Their phases are held together, so oscillators sharing a rate keep their relationship for as long as the device runs | Aurora, PolarNoise, Trails, Nebula, Fluid | — |
 | `map32` | Maps a value between ranges, clamped, with the fencepost handled once so the last column is never lost | FreqMatrix, FreqSaws, GEQ, GEQ3D, Spectrum, StarField | — |
 | `hashInt` | Hashes position and time to a random-looking but reproducible value, so devices agree without exchanging anything | Dissolve, WaterRipple | — |
-| `peakHold` | Rises instantly to a new high then decays slowly — the falling peak dot every VU meter has | Spectrum | — |
+| `peakHold` | Rises instantly to a new high then decays slowly, the falling peak dot every VU meter has | Spectrum | — |
 | `smoothFollow` | Moves a value a fraction of the way toward its target each frame, so it stops jittering | Spectrum | — |
 | `easeInOutQuad` | Accelerates from rest and decelerates to rest, so motion reads as deliberate rather than mechanical | Dissolve | — |
 | `easeInOutCubic`, `easeOutQuad` | The same family with a longer settle, and a fast-start curve that arrives and rests | *(no caller yet)* | — |
@@ -248,7 +231,7 @@ The framerate rule lives here too: everything in this group is driven by elapsed
 
 **Things that move under forces: sparks, rain, snow, smoke, confetti, debris, a swarm.**
 
-Anything that behaves like matter is the same handful of forces over the same state, and the part that differs between one look and another is *which* forces are applied and how particles are emitted — not the physics. So the state and the integrator live in [particles.h](../../../src/light/powerfunctions/particles.h) and the character stays with the effect.
+Anything that behaves like matter is the same handful of forces over the same state, and the part that differs between one look and another is *which* forces are applied and how particles are emitted, not the physics. So the state and the integrator live in [particles.h](../../../src/light/powerfunctions/particles.h) and the character stays with the effect.
 
 Storage is structure-of-arrays over the caller's own buffers, so a pass that touches only velocity walks only velocity, and the pool never allocates after `prepare()`. Positions are the same sub-pixel type `splat` takes, so a particle at x=3.5 lands half on each pixel instead of snapping.
 
@@ -268,7 +251,7 @@ A script reaches the same kernel through [the script vocabulary](moonlive.md#the
 |---|---|---|---|
 | `Pool` | The state: SoA positions, velocities, life, hue and optional size over caller-owned buffers. Owns nothing, allocates nothing | Fireworks, Ballpit, Particles | — |
 | `gravity`, `force`, `drag`, `attract` | The forces. Each is one pass over one array, so an effect pays only for the ones it uses | Fireworks, Ballpit | — |
-| `forceSmall` | A force too weak to move an integer velocity, accumulated until it does — what makes a light breeze read as inertia rather than as nothing | *(no caller yet)* | — |
+| `forceSmall` | A force too weak to move an integer velocity, accumulated until it does, what makes a light breeze read as inertia rather than as nothing | *(no caller yet)* | — |
 | `step` | Semi-implicit Euler: position integrates the already-updated velocity, which is what stays stable under a constant force | Fireworks, Ballpit, Particles | — |
 | `bounce`, `wrap`, `killOutside` | What happens at the walls: reflect with restitution, re-enter the opposite edge (snow, rain, marquee), or simply stop existing | Fireworks, Ballpit, Particles | — |
 | `collide` | Particles notice each other. The one non-linear part of the kernel, so it is opt-in | Ballpit | — |
@@ -280,9 +263,9 @@ A script reaches the same kernel through [the script vocabulary](moonlive.md#the
 
 ## Shaders
 
-**One function of (position, time) evaluated per pixel — the other way to write an effect.**
+**One function of (position, time) evaluated per pixel, the other way to write an effect.**
 
-Everything above draws *into* a grid: set this pixel, walk this line, move this row. A shader inverts that — it never draws anything, it answers a question. Given where a pixel is and what time it is, what color is it? The framework runs that function everywhere.
+Everything above draws *into* a grid: set this pixel, walk this line, move this row. A shader inverts that, it never draws anything, it answers a question. Given where a pixel is and what time it is, what color is it? The framework runs that function everywhere.
 
 That inversion is why shaders compose so freely. There is no state to keep in step and no order of operations to get right, so an effect is built by transforming the *coordinate* before answering: fold space and one shape becomes a thousand, rotate it and the whole design turns, displace it by a noise field and everything flows.
 
@@ -293,19 +276,19 @@ That inversion is why shaders compose so freely. There is no state to keep in st
 | Power function | What it does | Effects | Modifiers |
 |---|---|---|---|
 | `each` | The runner: supply one function of position and time, and it handles the loop, the coordinate mapping and the write | Truchet, Raymarch | — |
-| `uv` | Pixel to shader space, centred and scaled by the SHORT side — which is what keeps a circle circular on a non-square panel | Truchet, Raymarch | — |
+| `uv` | Pixel to shader space, centred and scaled by the SHORT side, which is what keeps a circle circular on a non-square panel | Truchet, Raymarch | — |
 | `clamp`, `mix`, `fract`, `step`, `smoothstep` | The five built-ins in essentially every shader. `fract` is the one that tiles a pattern; `smoothstep` is the one that anti-aliases an edge | Truchet | — |
 | `length`, `rotate` | Vector basics. Rotating the coordinate spins the entire design for one operation | Truchet | — |
-| `repeat`, `mirror` | Domain operators: fold space so one shape becomes a lattice. The objects do not multiply — the coordinate does the work | Truchet | — |
+| `repeat`, `mirror` | Domain operators: fold space so one shape becomes a lattice. The objects do not multiply, the coordinate does the work | Truchet | — |
 | `opUnion`, `opIntersect`, `opSubtract`, `opShell`, `opRound` | Combine two shapes into a third, which is how an SDF scene is composed rather than drawn | Truchet | — |
 | `sdRoundBox`, `sdPolygon` | Shapes beyond the circle/box/segment trio in [Signed distance fields](#signed-distance-fields) | *(no caller yet)* | — |
 | `cosPalette`, `mixColor` | A whole color ramp as twelve numbers instead of a table | *(no caller yet)* | — |
 
 </div>
 
-### Raymarching — one technique inside a shader
+### Raymarching: one technique inside a shader
 
-Raymarching is one technique a shader can use, for rendering 3D. A scene is described as a *function*: say how far the nearest surface is from any point, and the renderer walks a ray outward until it arrives. The world is arithmetic — geometry emerges from the distance function rather than being stored.
+Raymarching is one technique a shader can use, for rendering 3D. A scene is described as a *function*: say how far the nearest surface is from any point, and the renderer walks a ray outward until it arrives. The world is arithmetic, geometry emerges from the distance function rather than being stored.
 
 [raymarch.h](../../../src/light/powerfunctions/raymarch.h) is compiled only where the SoC declares a hardware FPU, because a raymarch is per-pixel float by nature. That gate is the one bounded exception to the integer-only render path, and it is a whole-header switch rather than a rule weakened in place. Everything in `shader.h` stays fixed point and runs everywhere.
 
@@ -314,10 +297,10 @@ Raymarching is one technique a shader can use, for rendering 3D. A scene is desc
 | Power function | What it does | Effects | Modifiers |
 |---|---|---|---|
 | `march` | Sphere tracing: walk a ray until it hits. Takes the scene as a callable, so that function *is* the world | Raymarch | — |
-| `normalAt` | The surface normal as the gradient of the distance field — which is why lighting works on a shape that was never modelled | Raymarch | — |
+| `normalAt` | The surface normal as the gradient of the distance field, which is why lighting works on a shape that was never modelled | Raymarch | — |
 | `sdSphere`, `sdBox`, `sdPlane`, `sdTorus` | 3D distance primitives, same sign contract as the 2D family | Raymarch | — |
 | `smin`, `opUnion`, `opIntersect`, `opSubtract`, `opRepeat` | The 3D operators. `smin` melts surfaces together; `opRepeat` tiles space into an endless lattice | Raymarch | — |
-| `Camera`, `diffuse` | Where the viewer stands and how a surface is lit — the parts every raymarch effect would otherwise re-derive | Raymarch | — |
+| `Camera`, `diffuse` | Where the viewer stands and how a surface is lit, the parts every raymarch effect would otherwise re-derive | Raymarch | — |
 
 </div>
 
@@ -331,7 +314,7 @@ Everything else writes; this reads. Once a frame can be sampled at an arbitrary
 
 | Power function | What it does | Effects | Modifiers |
 |---|---|---|---|
-| `sampleWrap` | Reads the grid as a texture at a sub-pixel coordinate, bilinear and wrapping — the primitive behind feedback, tunnels and zoom | Echo | — |
+| `sampleWrap` | Reads the grid as a texture at a sub-pixel coordinate, bilinear and wrapping, the primitive behind feedback, tunnels and zoom | Echo | — |
 | `combineMax` | Combines two colors by the brighter channel, so a trail brightens instead of averaging away | *(no caller yet)* | — |
 
 </div>
diff --git a/docs/moonmodules/light/supporting.md b/docs/moonmodules/light/supporting.md
index fadf9e7f..55b9bb94 100644
--- a/docs/moonmodules/light/supporting.md
+++ b/docs/moonmodules/light/supporting.md
@@ -1,16 +1,16 @@
 # Light supporting modules
 
-The light-domain machinery the catalog modules (effects, modifiers, layouts, drivers) lean on — not directly user-facing. Every row links to its generated technical page (the full API, from the `.h`) and its tests. Cross-cutting rationale that no single `.h` owns lives in the prose sections below the table.
+The light-domain machinery the catalog modules (effects, modifiers, layouts, drivers) lean on, not directly user-facing. Every row links to its generated technical page (the full API, from the `.h`) and its tests. Cross-cutting rationale that no single `.h` owns lives in the prose sections below the table.
 
 <a id="layer"></a>
 
 ### Layer
 
-One rendering layer — an effect writes into its buffer, modifiers transform the coordinate mapping, and the layer composites onto the shared output. The unit the render loop iterates.
+One rendering layer, an effect writes into its buffer, modifiers transform the coordinate mapping, and the layer composites onto the shared output. The unit the render loop iterates.
 
 <img src="../../assets/light/Layer.png" width="300" alt="Layer container with a child effect">
 
-- `blendMode` — how this layer composites onto the ones below (overwrite / alpha / additive).
+- `blendMode`: how this layer composites onto the ones below (overwrite / alpha / additive).
 
 Detail: [technical](moxygen/Layer.md)
 
@@ -20,7 +20,7 @@ Detail: [technical](moxygen/Layer.md)
 
 ### Effects
 
-The container of layers — composites them (blend mode + opacity per layer) into the final light buffer.
+The container of layers, composites them (blend mode + opacity per layer) into the final light buffer.
 
 <img src="../../assets/light/Effects.png" width="300" alt="Effects container">
 
@@ -32,7 +32,7 @@ Detail: [technical](moxygen/Effects.md)
 
 ### Layouts
 
-The container of layout modules — walks each layout's coordinates to build the physical light set the mapping consumes.
+The container of layout modules, walks each layout's coordinates to build the physical light set the mapping consumes.
 
 <img src="../../assets/light/Layouts.png" width="300" alt="Layouts container">
 
@@ -48,15 +48,15 @@ The container of driver modules, owning the shared buffer and the per-light outp
 
 <img src="../../assets/light/Drivers.png" width="300" alt="Drivers container with the on/off + brightness controls">
 
-- `on` — master power, on by default.
+- `on`: master power, on by default.
 
   Off scales the output to black while preserving `brightness`, so switching on restores the level. Every consumer drives it: the UI, infrared, the WLED app, Home Assistant and MQTT.
-- `brightness` — global output brightness, multiplied with each driver's `localBrightness`.
-- `palette` — the active palette effects sample from, built in or [scripted](moonlive.md).
-- `multicore` — run the output stage on the second core.
+- `brightness`: global output brightness, multiplied with each driver's `localBrightness`.
+- `palette`: the active palette effects sample from, built in or [scripted](moonlive.md).
+- `multicore`: run the output stage on the second core.
 
   On by default, and it falls back to single-core by itself when the extra frame buffer will not fit. The switch exists to compare the two.
-- `renderWait` — read-only: how long core 0 waited for core 1 at the frame boundary.
+- `renderWait`: read-only: how long core 0 waited for core 1 at the frame boundary.
 
   Near zero means render and output overlap well, and a large value means core 0 idles on a slow output stage. Shown only while `multicore` is on.
 
@@ -68,9 +68,9 @@ Detail: [technical](moxygen/Drivers.md)
 
 ### LightPresets
 
-The named channel wirings drivers reference — which channel carries Red, Green, Blue, White, or a fixture role like Pan/Tilt. Real fixtures ship read-only (the color orders, multi-channel pars, moving heads); add your own alongside them. A driver stores a preset's stable id, not its name, so renaming or reordering never breaks a reference.
+The named channel wirings drivers reference, which channel carries Red, Green, Blue, White, or a fixture role like Pan/Tilt. Real fixtures ship read-only (the color orders, multi-channel pars, moving heads); add your own alongside them. A driver stores a preset's stable id, not its name, so renaming or reordering never breaks a reference.
 
-- `presets` — the editable list of preset definitions, one row per preset.
+- `presets`: the editable list of preset definitions, one row per preset.
 
   A row carries a name, a channel count, and one role picker per channel. Built-in rows are read-only, and custom rows persist across a reboot.
 
@@ -78,7 +78,7 @@ Detail: [technical](moxygen/LightPresetsModule.md)
 
 ### Buffer
 
-Contiguous light-data buffer, shared between the layers that write it (effects) and the driver groups that read it. A raw `uint8_t*` so any channel layout fits — RGB, RGBW, multi-channel DMX.
+Contiguous light-data buffer, shared between the layers that write it (effects) and the driver groups that read it. A raw `uint8_t*` so any channel layout fits, RGB, RGBW, multi-channel DMX.
 
 Detail: [technical](moxygen/Buffer.md)
 
@@ -86,7 +86,7 @@ Detail: [technical](moxygen/Buffer.md)
 
 ### MappingLUT
 
-Maps the virtual grid to the physical sparse light set — a radius-4 sphere becomes its 210 real lights, not the 729-cell box. The lookup effects and the preview both consume.
+Maps the virtual grid to the physical sparse light set, a radius-4 sphere becomes its 210 real lights, not the 729-cell box. The lookup effects and the preview both consume.
 
 Detail: [technical](moxygen/MappingLUT.md)
 
@@ -116,31 +116,31 @@ Detail: [technical](moxygen/MoonLiveScript.md)
 
 ### Effect base
 
-The `EffectBase` class every effect derives from — the shared surface (buffer access, dimensions, the palette) an effect renders against.
+The `EffectBase` class every effect derives from, the shared surface (buffer access, dimensions, the palette) an effect renders against.
 
 Detail: [technical](moxygen/EffectBase.md)
 
 ### Modifier base
 
-The `ModifierBase` class every modifier derives from — transforms the coordinate mapping (mirror, rotate, multiply, …) a layer applies before rendering.
+The `ModifierBase` class every modifier derives from, transforms the coordinate mapping (mirror, rotate, multiply, …) a layer applies before rendering.
 
 Detail: [technical](moxygen/ModifierBase.md)
 
 ### Driver base
 
-The `DriverBase` class every driver derives from — the shared surface (the driver window, the source buffer, the output correction) a driver reads before sending its slice.
+The `DriverBase` class every driver derives from, the shared surface (the driver window, the source buffer, the output correction) a driver reads before sending its slice.
 
 Detail: [technical](moxygen/DriverBase.md)
 
 ### Layout base
 
-The `LayoutBase` class every layout derives from — the shared surface a layout implements to walk its coordinates into the physical light set the mapping consumes.
+The `LayoutBase` class every layout derives from, the shared surface a layout implements to walk its coordinates into the physical light set the mapping consumes.
 
 Detail: [technical](moxygen/LayoutBase.md)
 
 ### Slot encoder
 
-Turns lights into WS2812 bus words: each data bit becomes three bus slots (pulse start / data / tail), and the data slot is an 8×8 bit transpose — lanes in, bit-planes out. Shared by every parallel driver, and the render loop's measured hot spot.
+Turns lights into WS2812 bus words: each data bit becomes three bus slots (pulse start / data / tail), and the data slot is an 8×8 bit transpose, lanes in, bit-planes out. Shared by every parallel driver, and the render loop's measured hot spot.
 
 Detail: [technical](moxygen/ParallelSlots.md)
 
diff --git a/docs/reference/MIGRATING.md b/docs/reference/MIGRATING.md
index 7315f357..ebc21645 100644
--- a/docs/reference/MIGRATING.md
+++ b/docs/reference/MIGRATING.md
@@ -1,6 +1,6 @@
 # Migrating
 
-The log of **breaking changes** — what changed between versions, and the action to take.
+The log of **breaking changes**, what changed between versions, and the action to take.
 
 projectMM ships **no migration code**: the persistence layer is robust by default (an absent key keeps the control's default, a stale value clamps to the new bounds, an unknown key is ignored), which absorbs almost all schema drift with zero migration-specific code. The rare change that a robust reader *cannot* absorb is **documented here instead of migrated**. A patching framework is deferred rather than rejected: it becomes the right tool if breaking format changes get frequent enough that ad-hoc losses pile up (a rough bar: more than five across a few releases) and users hold persisted state too valuable to re-derive. At that point build the recognizable version-stamp plus ordered-patch-chain pattern, not a bespoke one and its rationale.
 
@@ -10,7 +10,7 @@ projectMM ships **no migration code**: the persistence layer is robust by defaul
 
 **MoonLive is exempt until it launches.** Nobody is running scripts on a device yet, so a break in the script language or its storage cannot strand anyone, and an entry here would describe an upgrade path no user can take. Its breaking changes are recorded in the commit and PR record instead. This exemption ends at the first release that ships MoonLive as a supported feature; from then it follows the same rule as everything else.
 
-**Action legend** — how much work an entry costs you:
+**Action legend**, how much work an entry costs you:
 
 | Action | Meaning |
 |---|---|
@@ -18,7 +18,7 @@ projectMM ships **no migration code**: the persistence layer is robust by defaul
 | *re-set a control* | One value resets to its default; set it again in the UI if you had changed it. |
 | *re-add a module* | The module vanishes from the tree on boot; add it again and re-enter its controls. |
 | *update a file* | An on-device file must be edited or replaced. |
-| *erase flash* | A full flash erase is required (the heaviest — a full reconfigure follows). |
+| *erase flash* | A full flash erase is required (the heaviest, a full reconfigure follows). |
 
 ---
 
@@ -79,104 +79,64 @@ Affects readers, not devices. A tutorial is a lesson for someone learning; putti
 **Action: re-set `floor` on a device whose microphone you had tuned.**
 Affects any device running the Audio module with a local microphone or line-in.
 
-`levels = automatic` is tuned with `floor` alone: how hard the learner levels, and how far it may
-lift a band, are constants rather than controls, because both act on a per-band range the
-conditioner has already normalized per rig, so one value serves every source.
+`levels = automatic` is tuned with `floor` alone: how hard the learner levels, and how far it may lift a band, are constants rather than controls, because both act on a per-band range the conditioner has already normalized per rig, so one value serves every source.
 
-`floor` is what changes meaning, and why re-setting it is worth a minute. It is now the **silence
-threshold** in both modes: below it a band reads zero and the learner does not learn from it. That
-is what stops a quiet room being amplified to full scale, but it also means a `floor` tuned under
-the old behavior can now gate audible sound. Raise it until a silent room reads still, then stop;
-there is no second knob to compensate with. `gain` remains manual-only and keeps its meaning.
+`floor` is what changes meaning, and why re-setting it is worth a minute. It is now the **silence threshold** in both modes: below it a band reads zero and the learner does not learn from it. That is what stops a quiet room being amplified to full scale, but it also means a `floor` tuned under the old behavior can now gate audible sound. Raise it until a silent room reads still, then stop; there is no second knob to compensate with. `gain` remains manual-only and keeps its meaning.
 
-Two behavior changes ride along and need no action. The spectrum now starts at 40 Hz rather than
-~11 Hz, dropping a first band that could only ever hold mains hum, DC drift and rumble. And
-AudioSpectrum's VU bar reads the raw level instead of the smoothed one, because it is the audio
-test instrument and wants maximum response; every other effect keeps the calm smoothed VU.
+Two behavior changes ride along and need no action. The spectrum now starts at 40 Hz rather than ~11 Hz, dropping a first band that could only ever hold mains hum, DC drift and rumble. And AudioSpectrum's VU bar reads the raw level instead of the smoothed one, because it is the audio test instrument and wants maximum response; every other effect keeps the calm smoothed VU.
 
 ### MoonBase serves the OTA routes under the application's names
 
 **Action: nothing on most devices; a serial flash on a MoonBase device updated from a browser.**
 Affects the 4 MB classic, `esp32-16mb` and the S3-Zero, the variants that carry MoonBase.
 
-MoonBase served `/install`, `/install-url`, `/boot-app`, `/last-url` and `/cancel` while the
-application served `/api/firmware/upload`, `/api/firmware/url` and `/api/firmware/moonbase`: two
-names for one operation, across images that a single browser page talks to in turn during one
-update. It now serves them under the application's names.
+MoonBase served `/install`, `/install-url`, `/boot-app`, `/last-url` and `/cancel` while the application served `/api/firmware/upload`, `/api/firmware/url` and `/api/firmware/moonbase`: two names for one operation, across images that a single browser page talks to in turn during one update. It now serves them under the application's names.
 
-The break is between the two images on a device, not between a device and its config. A device
-whose MoonBase predates this change still answers only the old names, so an updated application
-handing over to it leaves the browser calling routes that image does not have. The way through is
-the same as any MoonBase update: flash both images over serial once
-([building.md](../how-to/building.md#flashing-a-running-device-over-the-network)). A device flashed serially
-from this version on is consistent and needs nothing.
+The break is between the two images on a device, not between a device and its config. A device whose MoonBase predates this change still answers only the old names, so an updated application handing over to it leaves the browser calling routes that image does not have. The way through is the same as any MoonBase update: flash both images over serial once ([building.md](../how-to/building.md#flashing-a-running-device-over-the-network)). A device flashed serially from this version on is consistent and needs nothing.
 
 ### `soundReactive` is now `audioReactive`
 
 **Action: re-set one control.** Affects Fish Tank, Flying Toasters, Pacman, Pong, Space Invaders,
 Sprite Fountain and MovingHead, if you had turned the control on.
 
-One name for one thing: the service is `AudioService`, the frame is `AudioFrame`, the effects are
-audio-reactive. The control that made a sprite effect follow the music was the last place still
-calling it sound, so it is renamed rather than left as the odd one out.
+One name for one thing: the service is `AudioService`, the frame is `AudioFrame`, the effects are audio-reactive. The control that made a sprite effect follow the music was the last place still calling it sound, so it is renamed rather than left as the odd one out.
 
-A restored config maps the old name to the new one and carries its value. On a device upgraded in
-place the control returns to its default (off); switch it back on where you had it.
+A restored config maps the old name to the new one and carries its value. On a device upgraded in place the control returns to its default (off); switch it back on where you had it.
 
 ### AudioVolume is gone
 
 **Action: pick another effect.** Affects any device with an AudioVolume effect on a layer.
 
-It drew one bar from the audio level, which every audio-reactive effect does as a side effect of
-what it actually draws. There is no successor to map it onto, so a restored config carrying an
-`AudioVolumeEffect` node finds no such type and the layer comes up without it. `GEQ` is the nearest
-thing if a literal meter is what you want.
+It drew one bar from the audio level, which every audio-reactive effect does as a side effect of what it actually draws. There is no successor to map it onto, so a restored config carrying an `AudioVolumeEffect` node finds no such type and the layer comes up without it. `GEQ` is the nearest thing if a literal meter is what you want.
 
 ### The Firmware card describes one image at a time
 
 **Action: none.** Affects nothing a user has set: every control involved is read-only.
 
-`firmwarePartition` is now `partition`, and `update_pct` is gone (an install's progress belongs in
-the overlay the UI raises while it runs, not in a row that sits at zero for the life of a device
-that is not mid-install). Where a device carries two images, a new `image` control selects whether
-those rows describe the running app or MoonBase in the factory slot.
+`firmwarePartition` is now `partition`, and `update_pct` is gone (an install's progress belongs in the overlay the UI raises while it runs, not in a row that sits at zero for the life of a device that is not mid-install). Where a device carries two images, a new `image` control selects whether those rows describe the running app or MoonBase in the factory slot.
 
 ### Noise2D is gone; Noise renders it
 
 **Action: re-set one control.** Affects any device with a Noise2D effect on a layer.
 
-The two noise effects were one effect with two names: `Noise` is `Dim::D3` and draws the identical
-field on a panel, so the 2D variant earned nothing. A restored config maps `Noise2DEffect` to
-`NoiseEffect` and carries `scale` across.
+The two noise effects were one effect with two names: `Noise` is `Dim::D3` and draws the identical field on a panel, so the 2D variant earned nothing. A restored config maps `Noise2DEffect` to `NoiseEffect` and carries `scale` across.
 
-What does not carry is `speed`. Noise2D took a 0..15 divisor of its own; Noise takes its rate from
-`bpm` on the shared beat clock, so there is no value to map onto. Set `bpm` to taste after
-restoring.
+What does not carry is `speed`. Noise2D took a 0..15 divisor of its own; Noise takes its rate from `bpm` on the shared beat clock, so there is no value to map onto. Set `bpm` to taste after restoring.
 
 ### Infrared is a list of learned rows, and the remote must be re-learned
 
 **Action: re-learn the remote.** Affects any device with a configured infrared service.
 
-`IrService` becomes `InfraredService`, rebuilt around rows: a row learns a code and points it at any
-`Module.control`, where the old module carried five fixed actions (`code on/off`, `code brightness
-up`, and so on) each bound to one predetermined behavior. The module itself carries over through
-[migrate.js](https://github.com/MoonModules/projectMM/blob/main/src/ui/migrate.js)'s type map, so it
-does not vanish from the tree, but the codes it held have no equivalent: a learned code used to be a
-control's value, and is now a row. Press the remote's keys again against the rows you want.
+`IrService` becomes `InfraredService`, rebuilt around rows: a row learns a code and points it at any `Module.control`, where the old module carried five fixed actions (`code on/off`, `code brightness up`, and so on) each bound to one predetermined behavior. The module itself carries over through [migrate.js](https://github.com/MoonModules/projectMM/blob/main/src/ui/migrate.js)'s type map, so it does not vanish from the tree, but the codes it held have no equivalent: a learned code used to be a control's value, and is now a row. Press the remote's keys again against the rows you want.
 
-Restoring a backup taken before the change reports the rename and flags the module for review rather
-than silently dropping it. A device upgraded WITHOUT restoring a backup keeps its infrared module
-and loses the codes.
+Restoring a backup taken before the change reports the rename and flags the module for review rather than silently dropping it. A device upgraded WITHOUT restoring a backup keeps its infrared module and loses the codes.
 
 ### The desktop build keeps its files in `build/fs`, not `build`
 
 **Action: move your data, or lose your settings.** Affects the DESKTOP build only, and only a
 developer running it from a repository checkout; devices are unaffected.
 
-A desktop install used the build directory itself as the device's filesystem, so the File Manager's
-root listed CMake caches, object archives and every ESP32 variant's build folder alongside the four
-directories a device actually has. It now roots at `build/fs`, so what the desktop shows is what a
-board shows.
+A desktop install used the build directory itself as the device's filesystem, so the File Manager's root listed CMake caches, object archives and every ESP32 variant's build folder alongside the four directories a device actually has. It now roots at `build/fs`, so what the desktop shows is what a board shows.
 
 An existing checkout starts with an empty-looking device, because its `.config` is one level up.
 Move what you want to keep:
@@ -186,38 +146,23 @@ mkdir -p build/fs
 mv build/.config build/moonlive build/.hls build/fs/ 2>/dev/null
 ```
 
-Nothing is deleted if you skip this: the old directories stay where they are, and the device simply
-starts fresh. `MM_DATA_DIR` still overrides the location, and a packaged desktop install (which uses
-the per-user data directory) is unchanged.
+Nothing is deleted if you skip this: the old directories stay where they are, and the device simply starts fresh. `MM_DATA_DIR` still overrides the location, and a packaged desktop install (which uses the per-user data directory) is unchanged.
 
 ### projectMM no longer appears in WLED apps by default
 
-Device discovery now announces on the multicast group `239.255.77.77` and, by default, **not** on
-the broadcast address WLED apps and devices browse. A projectMM device therefore stops showing up
-in them until you turn on `wledCompatible` in the Devices module.
+Device discovery now announces on the multicast group `239.255.77.77` and, by default, **not** on the broadcast address WLED apps and devices browse. A projectMM device therefore stops showing up in them until you turn on `wledCompatible` in the Devices module.
 
-projectMM devices still find each other either way: presence always goes to the group and every
-device always joins it, so a fleet can mix the setting freely.
+projectMM devices still find each other either way: presence always goes to the group and every device always joins it, so a fleet can mix the setting freely.
 
-The reason for the default: a broadcast at discovery cadence makes every phone, printer and laptop
-on the LAN take an interrupt and parse a packet none of them want. Multicast reaches only the
-devices that joined the group. See
-[multicast and IGMP snooping](../explanation/architecture/moonlight.md#multicast-and-igmp-snooping) for when that saving is
-real (a switch that snoops) and when it is not.
+The reason for the default: a broadcast at discovery cadence makes every phone, printer and laptop on the LAN take an interrupt and parse a packet none of them want. Multicast reaches only the devices that joined the group. See [multicast and IGMP snooping](../explanation/architecture/moonlight.md#multicast-and-igmp-snooping) for when that saving is real (a switch that snoops) and when it is not.
 
 ### A light preset's Dimmer channel is now driven
 
-A preset that declares a `Dimmer` role previously left that channel at 0, because nothing ever
-wrote it: `Correction` resolved only the color roles. A fixture on such a preset therefore emitted
-nothing at all, whatever its color channels said. The shipped `IRGB` preset ("CH1 master
-intensity") could never light a fixture.
+A preset that declares a `Dimmer` role previously left that channel at 0, because nothing ever wrote it: `Correction` resolved only the color roles. A fixture on such a preset therefore emitted nothing at all, whatever its color channels said. The shipped `IRGB` preset ("CH1 master intensity") could never light a fixture.
 
-The dimmer is now held open (255) every frame, with per-light brightness staying in the color
-values as before. **If you drive a fixture on `IRGB` or another dimmer-carrying preset, it will
-light up where it previously stayed dark.** Nothing to change; the previous behavior was a defect.
+The dimmer is now held open (255) every frame, with per-light brightness staying in the color values as before. **If you drive a fixture on `IRGB` or another dimmer-carrying preset, it will light up where it previously stayed dark.** Nothing to change; the previous behavior was a defect.
 
-Routing brightness to the dimmer channel rather than holding it open is the better model and is
-[backlogged](../work/future/backlog-light.md), so this value will change again.
+Routing brightness to the dimmer channel rather than holding it open is the better model and is [backlogged](../work/future/backlog-light.md), so this value will change again.
 
 
 ### esp32-16mb moves to the MoonBase partition table (2026-08-28)
@@ -225,12 +170,7 @@ Routing brightness to the dimmer channel rather than holding it open is the bett
 **Action: erase flash** (USB re-flash). Back up first (File Manager, or the installer's
 bookmarklet on older firmware); restore after the install brings WiFi, config and scripts back.
 
-`esp32-16mb` replaces its dual-OTA layout with
-[MoonBase](../explanation/architecture/moonbase.md), the same trade the 4 MB variants
-made in the entry below, taken here by choice rather than necessity: the second app slot was
-idle except during an update, so the filesystem grows 7168 to 11264 KB and the device gains
-MoonBase's stronger recovery story (a power cut mid-install boots MoonBase and the user retries
-over the network). One app slot remains, at its full 4096 KB.
+`esp32-16mb` replaces its dual-OTA layout with [MoonBase](../explanation/architecture/moonbase.md), the same trade the 4 MB variants made in the entry below, taken here by choice rather than necessity: the second app slot was idle except during an update, so the filesystem grows 7168 to 11264 KB and the device gains MoonBase's stronger recovery story (a power cut mid-install boots MoonBase and the user retries over the network). One app slot remains, at its full 4096 KB.
 
 Every partition moves, so the existing filesystem volume is not where the new table looks:
 without a backup, WiFi credentials, module config and scripts all re-enter through provisioning.
@@ -241,13 +181,7 @@ A partition table only changes over USB, so an OTA update leaves a device on the
 **Action: erase flash** (USB re-flash). Back up first (File Manager ⤓, or the installer's
 bookmarklet on older firmware); restore after the install brings WiFi, config and scripts back.
 
-The 4 MB variants (`esp32`, `esp32-wrover`, `esp32-eth`) replace the dual-OTA layout with
-[MoonBase](../explanation/architecture/moonbase.md): the app slot grows
-1856 → 2496 KB and the filesystem 256 → 548 KB, but the filesystem moves (0x3B0000 → 0x360000),
-so the existing volume is not where the new table looks; without a backup, WiFi credentials,
-module config and scripts all re-enter through provisioning. A partition table only changes over USB: a device
-still on the old table keeps OTA-updating *within* that table for as long as the app fits its
-1856 KB slot; the web installer is the migration path. 8/16 MB boards are unaffected.
+The 4 MB variants (`esp32`, `esp32-wrover`, `esp32-eth`) replace the dual-OTA layout with [MoonBase](../explanation/architecture/moonbase.md): the app slot grows 1856 → 2496 KB and the filesystem 256 → 548 KB, but the filesystem moves (0x3B0000 → 0x360000), so the existing volume is not where the new table looks; without a backup, WiFi credentials, module config and scripts all re-enter through provisioning. A partition table only changes over USB: a device still on the old table keeps OTA-updating *within* that table for as long as the app fits its 1856 KB slot; the web installer is the migration path. 8/16 MB boards are unaffected.
 
 ### PreviewDriver's `fps` becomes `targetFps`, and now trades resolution (2026-08-25)
 
@@ -261,32 +195,22 @@ The control is renamed and its meaning changed, so the rename is the point rathe
 
 ### A module declares every control with `addControl` (2026-08-24)
 
-`addUint8`, `addUint16`, `addInt16`, `addInt32` and `addBool` are replaced by one overloaded
-`addControl(name, variable, min, max)`. The widget follows the variable's own type, which the
-compiler already knows, so the name no longer repeats a width the declaration states:
+`addUint8`, `addUint16`, `addInt16`, `addInt32` and `addBool` are replaced by one overloaded `addControl(name, variable, min, max)`. The widget follows the variable's own type, which the compiler already knows, so the name no longer repeats a width the declaration states:
 
 ```cpp
 controls_.addUint8("speed", speed_, 1, 255);     // before
 controls_.addControl("speed", speed_, 1, 255);   // after
 ```
 
-This is the same call a MoonLive script makes, which is the point: someone who has written a
-script can read a compiled module, and someone who has read a module can write a script.
+This is the same call a MoonLive script makes, which is the point: someone who has written a script can read a compiled module, and someone who has read a module can write a script.
 
-The **widget-specific** adders keep their names — `addPin`, `addSelect`, `addPalette`, `addText`,
-`addTextArea`, `addFilePath`, `addPassword`, `addIPv4`, `addReadOnly`, `addReadOnlyInt`,
-`addProgress`, `addList`, `addButton`. Those name a widget rather than a width, and the intent is
-not recoverable from the C++ type: `uint8_t` backs a slider, a dropdown *and* a palette picker, and
-an `int8_t` silently becoming a Pin would register as a claimed GPIO in the pin map. `addControl`
-on an `int8_t` is deliberately deleted, with a diagnostic naming the two real options.
+The **widget-specific** adders keep their names, `addPin`, `addSelect`, `addPalette`, `addText`, `addTextArea`, `addFilePath`, `addPassword`, `addIPv4`, `addReadOnly`, `addReadOnlyInt`, `addProgress`, `addList`, `addButton`. Those name a widget rather than a width, and the intent is not recoverable from the C++ type: `uint8_t` backs a slider, a dropdown *and* a palette picker, and an `int8_t` silently becoming a Pin would register as a claimed GPIO in the pin map. `addControl` on an `int8_t` is deliberately deleted, with a diagnostic naming the two real options.
 
 **Action: *nothing* for a device.** No control name, type, range, wire format or persisted value
-changes — a renamed call produces a byte-identical descriptor, which is why nothing on the device
-can notice.
+changes, a renamed call produces a byte-identical descriptor, which is why nothing on the device can notice.
 
 **Action for a third-party module: *recompile*.** Rename the five calls to `addControl`; the
-arguments are unchanged. A missed one is a compile error, never a silent behavior change: the
-overloads bind by exact reference type, so a call that compiles produces the widget it always did.
+arguments are unchanged. A missed one is a compile error, never a silent behavior change: the overloads bind by exact reference type, so a call that compiles produces the widget it always did.
 
 
 ### Desktop settings move to a per-user directory (2026-08-23)
@@ -304,7 +228,7 @@ ESP32 is unaffected: LittleFS mounts at a fixed partition and never used this pa
 
 ### The `Layers` container is renamed to `Effects` (2026-08-08)
 
-The three top-level light containers are now **Layouts, Effects, Drivers** — L.E.D. The old name sat one character from its own child (`Layers` holding `Layer`s) and read as a near-twin of `Layouts`, which is the pair a newcomer actually has to tell apart. The tree is unchanged in shape: `Effects` → `Layer`s → effects and modifiers.
+The three top-level light containers are now **Layouts, Effects, Drivers**, L.E.D. The old name sat one character from its own child (`Layers` holding `Layer`s) and read as a near-twin of `Layouts`, which is the pair a newcomer actually has to tell apart. The tree is unchanged in shape: `Effects` → `Layer`s → effects and modifiers.
 
 **Action: *re-add a module* and *re-save presets*.**
 
@@ -322,17 +246,17 @@ The child `Layer` keeps its name, as does everything under it.
 
 ### The `peripheral` options are renamed to name the peripheral, not the bus protocol (2026-07-30)
 
-The `peripheral` dropdown no longer says `i80` / `MoonI80`. "i80" is the Intel 8080 bus shape `esp_lcd` speaks — it is not a peripheral any ESP32 datasheet lists, and it matched nothing a user could look up: on the classic ESP32 that backend **is the I2S peripheral**, on the S3/P4/S31 it is the **LCD** peripheral. The new labels name the silicon block plus who drives it, which is the actual choice being made.
+The `peripheral` dropdown no longer says `i80` / `MoonI80`. "i80" is the Intel 8080 bus shape `esp_lcd` speaks, it is not a peripheral any ESP32 datasheet lists, and it matched nothing a user could look up: on the classic ESP32 that backend **is the I2S peripheral**, on the S3/P4/S31 it is the **LCD** peripheral. The new labels name the silicon block plus who drives it, which is the actual choice being made.
 
 | Old | New (classic ESP32) | New (S3 / P4 / S31) |
 |---|---|---|
 | `i80` | `I2S-IDF` | `LCD-IDF` |
-| `MoonI80` | — (not available) | `LCD-MM` |
-| `Parlio` | — | `Parlio` (unchanged — it *is* the peripheral's name) |
+| `MoonI80` |, (not available) | `LCD-MM` |
+| `Parlio` | — | `Parlio` (unchanged, it *is* the peripheral's name) |
 
 `-IDF` = driven through ESP-IDF's `esp_lcd`; `-MM` = driven by our own GDMA layer below it, which is what buys the streaming ring and the 74HCT595 pin expander.
 
-**Action: re-set the `peripheral` control** — but only on a device that already holds a persisted parallel driver AND had a non-default peripheral selected. The stored string no longer matches any option, so the loader falls back to the board's default backend; if that was already your choice, nothing changes. The web installer's board catalog ships the new names, so a fresh install or catalog re-inject is correct without action.
+**Action: re-set the `peripheral` control**, but only on a device that already holds a persisted parallel driver AND had a non-default peripheral selected. The stored string no longer matches any option, so the loader falls back to the board's default backend; if that was already your choice, nothing changes. The web installer's board catalog ships the new names, so a fresh install or catalog re-inject is correct without action.
 
 ### The three parallel LED drivers merge into one `ParallelLedDriver` with a `peripheral` selector (2026-07-23)
 
@@ -340,15 +264,15 @@ The `peripheral` dropdown no longer says `i80` / `MoonI80`. "i80" is the Intel 8
 
 | Old registered type | New |
 |---|---|
-| `MultiPinLedDriver` | `ParallelLedDriver` + `peripheral` = `i80` (esp_lcd: LCD_CAM on S3/P4, I2S on classic) — renamed again below |
-| `MoonLedDriver` | `ParallelLedDriver` + `peripheral` = `MoonI80` (own-GDMA below esp_lcd, LCD_CAM) — renamed again below |
+| `MultiPinLedDriver` | `ParallelLedDriver` + `peripheral` = `i80` (esp_lcd: LCD_CAM on S3/P4, I2S on classic), renamed again below |
+| `MoonLedDriver` | `ParallelLedDriver` + `peripheral` = `MoonI80` (own-GDMA below esp_lcd, LCD_CAM), renamed again below |
 | `ParlioLedDriver` | `ParallelLedDriver` + `peripheral` = `Parlio` (P4) |
 
-**Action: re-add the driver.** A persisted module whose type is one of the three old names no longer resolves (the type isn't registered), so the robust loader drops it on boot — the driver, and its pins/settings, vanish from the tree. Add a **Parallel LED** driver again, choose the `peripheral` your board uses (the same backend the old type named — see the table), and re-enter its `pins` / `ledsPerPin` plus whatever the chosen peripheral needs: `i80` has `clockPin`/`dcPin`, `MoonI80` has `clockPin` + the ring/expander controls, `Parlio` has no clock or DC pins at all. The web installer's board catalog already names the new type, so a fresh install or a catalog re-inject wires it correctly; only a device carrying an OLD persisted tree needs the manual re-add.
+**Action: re-add the driver.** A persisted module whose type is one of the three old names no longer resolves (the type isn't registered), so the robust loader drops it on boot, the driver, and its pins/settings, vanish from the tree. Add a **Parallel LED** driver again, choose the `peripheral` your board uses (the same backend the old type named, see the table), and re-enter its `pins` / `ledsPerPin` plus whatever the chosen peripheral needs: `i80` has `clockPin`/`dcPin`, `MoonI80` has `clockPin` + the ring/expander controls, `Parlio` has no clock or DC pins at all. The web installer's board catalog already names the new type, so a fresh install or a catalog re-inject wires it correctly; only a device carrying an OLD persisted tree needs the manual re-add.
 
 ### The per-driver `preset` control is renamed to `lightPreset` (2026-07-23)
 
-**Action: nothing** on-device (the saved value survives, see the `lightPreset` [persistence contract](../moonmodules/light/drivers.md#led-driver-details)). Only an external script or automation that POSTs the control by name (`/api/control` with `"control":"preset"`) must switch to `lightPreset`.
+**Action: nothing** on-device (the saved value survives, see the `lightPreset` [persistence contract](../moonmodules/light/drivers.md#parallel-led-details)). Only an external script or automation that POSTs the control by name (`/api/control` with `"control":"preset"`) must switch to `lightPreset`.
 
 ### `AudioService`: the `sync` control becomes `mode` + `send audio`, and `simulate` is renumbered (2026-07-22)
 
@@ -357,15 +281,15 @@ The audio module's identity is now a single `mode` control (Local audio / Receiv
 | Old | New |
 |---|---|
 | control `sync` (Select: `off`/`send`/`receive`) | `mode` (Select: `local audio`/`receive network`/`simulate`) + `send audio` (a switch, Local mode only) |
-| control `simulate` (Select, 5 options incl. a mic-fill-on-silence mode) | `simulate` (Select: 2 options) — used only when `mode` is Simulate |
+| control `simulate` (Select, 5 options incl. a mic-fill-on-silence mode) | `simulate` (Select: 2 options), used only when `mode` is Simulate |
 
 **Action: re-set `mode` (and `send audio`) if you had `sync` on `send` or `receive`; re-set `simulate` if you had chosen a non-default option.**
 
-`sync` and the old `simulate` value read as absent → ignored, so `mode` takes its default (**Local audio**) and `send audio` its default (**off**). A device that was on `sync=receive` therefore comes up as Local audio — set `mode` to Receive network again. One that broadcast (`sync=send`) comes up not broadcasting — turn `send audio` on. The five-option `simulate` collapsed to two, so a device on one of the dropped options (e.g. the mic-fill-on-silence mode, a removed capability) takes the new default; re-pick if needed. Receive network and every sync control exist only on network-capable targets.
+`sync` and the old `simulate` value read as absent → ignored, so `mode` takes its default (**Local audio**) and `send audio` its default (**off**). A device that was on `sync=receive` therefore comes up as Local audio, set `mode` to Receive network again. One that broadcast (`sync=send`) comes up not broadcasting, turn `send audio` on. The five-option `simulate` collapsed to two, so a device on one of the dropped options (e.g. the mic-fill-on-silence mode, a removed capability) takes the new default; re-pick if needed. Receive network and every sync control exist only on network-capable targets.
 
 ### `MoonLedDriver`: `forceRing` → `useRing`, and the ring's geometry is now settable (2026-07-17)
 
-The pin-expander path selector was a three-option Select (`auto` / `ring` / `wholeFrame`) named for a *diagnostic override*. The auto-router is gone — at the size the expander exists for (48 strands × 256 lights) a whole frame never fits internal DMA RAM, so "auto" had exactly one right answer while presenting itself as a choice, and its silent fallback hid which path was actually running. What remains is the honest question, as a switch:
+The pin-expander path selector was a three-option Select (`auto` / `ring` / `wholeFrame`) named for a *diagnostic override*. The auto-router is gone, at the size the expander exists for (48 strands × 256 lights) a whole frame never fits internal DMA RAM, so "auto" had exactly one right answer while presenting itself as a choice, and its silent fallback hid which path was actually running. What remains is the honest question, as a switch:
 
 | Old | New |
 |---|---|
@@ -375,9 +299,9 @@ The pin-expander path selector was a three-option Select (`auto` / `ring` / `who
 
 **Action: re-set `useRing` if you had `forceRing` on `wholeFrame`.**
 
-`forceRing` reads as absent → ignored, and `useRing` takes its default (**on**, the ring). A device that had explicitly selected whole-frame therefore comes up on the ring; flip `useRing` off to get it back. `ringRows`/`ringBufs` default to 16 and 12 — the geometry the driver effectively ran. (It shipped with a pool of 16, but 16 buffers never fit the S3's internal DMA heap, so the ring build failed its own fit check and the driver quietly fell back to whole-frame; 12 is what actually held. A config on the old defaults may therefore start *ringing* where it used to fall back.) They exist so the RAM / encode-overhead / interrupt-rate / lap-time trade-off can be swept on a live board rather than fixed at compile time.
+`forceRing` reads as absent → ignored, and `useRing` takes its default (**on**, the ring). A device that had explicitly selected whole-frame therefore comes up on the ring; flip `useRing` off to get it back. `ringRows`/`ringBufs` default to 16 and 12, the geometry the driver effectively ran. (It shipped with a pool of 16, but 16 buffers never fit the S3's internal DMA heap, so the ring build failed its own fit check and the driver quietly fell back to whole-frame; 12 is what actually held. A config on the old defaults may therefore start *ringing* where it used to fall back.) They exist so the RAM / encode-overhead / interrupt-rate / lap-time trade-off can be swept on a live board rather than fixed at compile time.
 
-### LED driver + control rename — a human-readable UI (2026-07-16)
+### LED driver + control rename: a human-readable UI (2026-07-16)
 
 The LED driver module types and several controls were renamed so the UI reads in plain language rather than peripheral jargon (the UI shows a control's name verbatim, so the name *is* the label).
 
@@ -392,7 +316,7 @@ The LED driver module types and several controls were renamed so the UI reads in
 
 **Action: re-add the module, then re-set `pinExpander` / `doubleBuffer` if you had changed them.**
 
-A device whose persisted config names the old module type loads a module type that no longer exists — the unknown type is ignored, so **the driver is absent from the tree on boot**. Re-add a **Parallel LED** driver (the single type the two later merged into — see the 2026-07-23 entry above for the `peripheral` value that matches the old `I80LedDriver` / `MoonI80LedDriver`) and re-enter its controls. Within a re-added driver, the two renamed *settable* controls (`pinExpander`, `doubleBuffer`) read as absent → they take their defaults (`pinExpander` off, `doubleBuffer` on); set them again if your board needs otherwise. `frameTime` and `renderWait` are read-only KPIs — nothing to restore.
+A device whose persisted config names the old module type loads a module type that no longer exists, the unknown type is ignored, so **the driver is absent from the tree on boot**. Re-add a **Parallel LED** driver (the single type the two later merged into, see the 2026-07-23 entry above for the `peripheral` value that matches the old `I80LedDriver` / `MoonI80LedDriver`) and re-enter its controls. Within a re-added driver, the two renamed *settable* controls (`pinExpander`, `doubleBuffer`) read as absent → they take their defaults (`pinExpander` off, `doubleBuffer` on); set them again if your board needs otherwise. `frameTime` and `renderWait` are read-only KPIs, nothing to restore.
 
 This rename left `RmtLedDriver` untouched, and `ParlioLedDriver` untouched *at the time*; the later 2026-07-23 entry above then merges `ParlioLedDriver` into `ParallelLedDriver` along with the other two. The `pins` / `ledsPerPin` / `clockPin` / `latchPin` / `loopback*` controls are unchanged by this rename.
 
diff --git a/docs/reference/hardware/control-surfaces.md b/docs/reference/hardware/control-surfaces.md
index 14ed8c01..01a49525 100644
--- a/docs/reference/hardware/control-surfaces.md
+++ b/docs/reference/hardware/control-surfaces.md
@@ -1,14 +1,10 @@
 # Control surfaces: hardware reference
 
-What projectMM needs to know about the physical desks on the bench, so a control-ingest plan can be
-written from facts rather than from a product page. A desk here is a candidate source for
-[ControlModule](../../moonmodules/core/system.md#control)'s pads, encoders and faders, which were laid out to
-match this class of hardware in the first place.
+What projectMM needs to know about the physical desks on the bench, so a control-ingest plan can be written from facts rather than from a product page. A desk here is a candidate source for [ControlModule](../../moonmodules/core/system.md#control)'s pads, encoders and faders, which were laid out to match this class of hardware in the first place.
 
 **The headline, because it contradicts the obvious assumption:** neither desk speaks OSC. Both are
 **Mackie Control** surfaces. OSC is the right protocol for the wider ecosystem (Resolume,
-TouchDesigner, TouchOSC, DIY Arduino rigs) and is planned on that basis, but it does not reach
-these two. See [the OSC plan](../../work/present/Plan-20260829%20-%20OSC%20control%20ingest.md).
+TouchDesigner, TouchOSC, DIY Arduino rigs) and is planned on that basis, but it does not reach these two. See [the OSC plan](../../work/present/Plan-20260829%20-%20OSC%20control%20ingest.md).
 
 ## Behringer X-Touch (Universal)
 
@@ -27,10 +23,7 @@ The larger of the two, and the only one with a network port.
 | Not supported | OSC |
 
 **The Ethernet port carries RTP-MIDI, not OSC.** RTP-MIDI (RFC 6295, UDP port 5004) is MIDI
-tunnelled over a network with a session layer: an invitation handshake, synchronisation, and a
-journal so a dropped packet can be recovered rather than losing a note. It is a real protocol to
-implement, not a framing detail, which is what separates "reach this desk over the LAN" from
-"reach it over USB".
+tunnelled over a network with a session layer: an invitation handshake, synchronization, and a journal so a dropped packet can be recovered rather than losing a note. It is a real protocol to implement, not a framing detail, which is what separates "reach this desk over the LAN" from "reach it over USB".
 
 ## iCON QCon Pro G2
 
@@ -52,8 +45,7 @@ host support and a machine in the rack is the practical answer today.
 
 ## What Mackie Control looks like on the wire
 
-Enough to judge the size of the job. MCU is ordinary MIDI carrying agreed meanings, so a parser is
-small; the work is in the semantics and the feedback, not the bytes.
+Enough to judge the size of the job. MCU is ordinary MIDI carrying agreed meanings, so a parser is small; the work is in the semantics and the feedback, not the bytes.
 
 | element | encoding |
 |---|---|
@@ -65,10 +57,7 @@ small; the work is in the semantics and the feedback, not the bytes.
 | Scribble strips | SysEx, `0x12` after the header, then the text |
 
 **It is bidirectional by nature, and that is the point of the hardware.** The motors only move
-because the host sends fader positions back; the scribble strips only show anything because the
-host writes them. A projectMM implementation that only *reads* the desk would work, but would
-waste what makes these desks worth owning: a preset change should move the faders and relabel the
-strips.
+because the host sends fader positions back; the scribble strips only show anything because the host writes them. A projectMM implementation that only *reads* the desk would work, but would waste what makes these desks worth owning: a preset change should move the faders and relabel the strips.
 
 ## What this means for projectMM
 
@@ -79,6 +68,4 @@ Three routes, in increasing cost:
 2. **A bridge.** An existing MCU-to-OSC translator on a laptop turns either desk into an OSC
    source. No firmware work; costs a machine in the rack, which a festival podium may already have.
 3. **RTP-MIDI + MCU in firmware.** Reaches the X-Touch over Ethernet with no host machine, and is
-   the only route that drives the motors from projectMM directly. Needs RFC 6295 (session
-   handshake, journalling) plus the MCU semantics above, both directions. Does not help the QCon,
-   which has no network port.
+   the only route that drives the motors from projectMM directly. Needs RFC 6295 (session handshake, journalling) plus the MCU semantics above, both directions. Does not help the QCon, which has no network port.
diff --git a/docs/reference/hardware/esp32-s31-coreboard.md b/docs/reference/hardware/esp32-s31-coreboard.md
index b93b50c9..a482cd2c 100644
--- a/docs/reference/hardware/esp32-s31-coreboard.md
+++ b/docs/reference/hardware/esp32-s31-coreboard.md
@@ -1,8 +1,6 @@
 # ESP32-S31 Function-CoreBoard-1 hardware reference
 
-Pin maps and onboard features for the Espressif **ESP32-S31 Function-CoreBoard-1**, read from the
-official schematic so projectMM work (Ethernet, audio, SD, USB-host) reads this instead of
-re-scraping the PDF. The board is the bench S31 (`esp32s31` firmware).
+Pin maps and onboard features for the Espressif **ESP32-S31 Function-CoreBoard-1**, read from the official schematic so projectMM work (Ethernet, audio, SD, USB-host) reads this instead of re-scraping the PDF. The board is the bench S31 (`esp32s31` firmware).
 
 **Sources**
 - Schematic (rev C, 2026-05-13): <https://dl.espressif.com/schematics/esp32-s31-function-coreboard-1-schematics.pdf>
@@ -11,8 +9,7 @@ re-scraping the PDF. The board is the bench S31 (`esp32s31` firmware).
 - Module (WROOM-3): <https://documentation.espressif.com/esp32-s31-wroom-3_datasheet_en.pdf>
 - User guide: <https://docs.espressif.com/projects/esp-dev-kits/en/latest/esp32s31/esp32-s31-function-coreboard-1/user_guide.html>
 
-The module is **ESP32-S31-WROOM-3**: RISC-V dual-core (≤300 MHz), Wi-Fi 6, BT, IEEE 802.15.4,
-on-chip 1 Gbps EMAC. It shares the MoonLive RISC-V backend with the ESP32-P4.
+The module is **ESP32-S31-WROOM-3**: RISC-V dual-core (≤300 MHz), Wi-Fi 6, BT, IEEE 802.15.4, on-chip 1 Gbps EMAC. It shares the MoonLive RISC-V backend with the ESP32-P4.
 
 ## Audio (ES8311 codec)
 
@@ -34,13 +31,11 @@ The onboard electret mic (J6) and speaker connect through an **ES8311 mono codec
 > `ESP_I2C_SCL` net labels suggest. Bench-confirmed: the [I2cScanModule](../../moonmodules/core/moxygen/I2cScanModule.md)
 > (sda=51, scl=50 in the S31 catalog entry) finds the ES8311 ACK at 0x18; with 50/51 nothing
 > ACKs. The other audio pins match the schematic + the chip's GPIO table (all of GPIO50–57 are
-> plain I/O GPIOs routed through the matrix — no special-function conflict).
+> plain I/O GPIOs routed through the matrix, no special-function conflict).
 
 - **ES8311 I2C address: `0x18`** (the default; set by the `CE` pin tie).
 - Driven by Espressif's **`esp_codec_dev`** managed component (the ES8311 driver). The codec
-  needs **MCLK running before it answers I2C**, and `es8311_codec_cfg.mclk_div` must be set
-  (256, the standard ratio) or `open` fails "unable to configure sample rate". So AudioService
-  brings up the I2S channel (which drives MCLK on GPIO52) **before** the codec I2C config.
+  needs **MCLK running before it answers I2C**, and `es8311_codec_cfg.mclk_div` must be set (256, the standard ratio) or `open` fails "unable to configure sample rate". So AudioService brings up the I2S channel (which drives MCLK on GPIO52) **before** the codec I2C config.
 - **Mic-only path** (audio-reactive input) needs MCLK/SCLK/LRCK + ASDOUT (record) + the I2C bus.
   The speaker path (DSDIN + PA_CTRL) is output, a separate capability.
 
@@ -48,9 +43,7 @@ The onboard electret mic (J6) and speaker connect through an **ES8311 mono codec
 
 On-chip EMAC → **YT8531** (Motorcomm) PHY (U8) → RJ45, **RGMII** with a 25 MHz crystal (Y2).
 
-Pin map — **bench-verified** (link + DHCP confirmed on the CoreBoard). The RGMII data + clock
-GPIOs are the chip's fixed IO_MUX pads (the only ones the EMAC accepts; from IDF's
-`esp32s31/emac_periph.c`); MDC/MDIO are IDF's S31 SMI defaults (`ETH_ESP32_EMAC_DEFAULT_CONFIG`):
+Pin map, **bench-verified** (link + DHCP confirmed on the CoreBoard). The RGMII data + clock GPIOs are the chip's fixed IO_MUX pads (the only ones the EMAC accepts; from IDF's `esp32s31/emac_periph.c`); MDC/MDIO are IDF's S31 SMI defaults (`ETH_ESP32_EMAC_DEFAULT_CONFIG`):
 
 | Signal | GPIO | | Signal | GPIO |
 |---|---|---|---|---|
@@ -74,27 +67,21 @@ GPIOs are the chip's fixed IO_MUX pads (the only ones the EMAC accepts; from IDF
 >
 > **RGMII 125 MHz Tx clock = APLL.** The EMAC needs a clean 125 MHz Tx clock; the default MPLL is
 > owned by PSRAM (400 MHz, no integer path to 125) and CPLL can't hit it on the 40 MHz XTAL grid, so
-> `sdkconfig.defaults.esp32s31` sets `CONFIG_ETH_EMAC_RGMII_TX_CLK_SRC_APLL` — the fractional PLL
+> `sdkconfig.defaults.esp32s31` sets `CONFIG_ETH_EMAC_RGMII_TX_CLK_SRC_APLL`, the fractional PLL
 > synthesises 125 MHz exactly.
 
 ## Other onboard features
 
-- **Addressable RGB LED** — WS2812 (D7) on **GPIO60**. *(Driven today: the catalog's S31
-  `RmtLedDriver` uses `pins: "60"`, with `loopbackTxPin: 48` / `loopbackRxPin: 47` for the
-  self-test — see [§ Free GPIO](#free-gpio-for-user-peripherals-led-strips-loopback).)*
-- **SD card slot** — SD_D0–D3 / SD_CLK / SD_CMD on the module's SDIO pins (GPIO20–25 per the user
-  guide). Note: `SOC_SDMMC_SUPPORTED` is **absent** in the S31 soc-caps, so the slot is likely
-  SPI-mode (GPSPI) rather than the SDMMC peripheral — confirm before relying on it.
-- **USB-A host** — USB 2.0 high-speed host port (5 V, 0.5 A limit).
-- **USB-C ×2** — one is USB Serial/JTAG (native), one is a USB-to-UART bridge (CP2102N default, or
+- **Addressable RGB LED**: WS2812 (D7) on **GPIO60**. *(Driven today: the catalog's S31
+  `RmtLedDriver` uses `pins: "60"`, with `loopbackTxPin: 48` / `loopbackRxPin: 47` for the self-test, see [§ Free GPIO](#free-gpio-for-user-peripherals-led-strips-loopback).)*
+- **SD card slot**: SD_D0–D3 / SD_CLK / SD_CMD on the module's SDIO pins (GPIO20–25 per the user
+  guide). Note: `SOC_SDMMC_SUPPORTED` is **absent** in the S31 soc-caps, so the slot is likely SPI-mode (GPSPI) rather than the SDMMC peripheral, confirm before relying on it.
+- **USB-A host**: USB 2.0 high-speed host port (5 V, 0.5 A limit).
+- **USB-C ×2**: one is USB Serial/JTAG (native), one is a USB-to-UART bridge (CP2102N default, or
   CH9102X with the alternate BOM). Auto-download via DTR/RTS → EN/BOOT.
-- **Buttons** — BOOT (GPIO61), RESET (EN).
-- **J5 — power/enable jumper** (a 2-pin header next to the 5V→3.3V DC/DC converter U4 and the 3.3V
-  power-on LED D11). **Leave it installed.** With J5 removed the board is *half-powered*: the CP2102N
-  runs off USB VBUS so its port still enumerates, but the ESP32-S31's 3.3V/EN rail is incomplete and
-  the chip drives nothing — you get a serial port that opens but zero bytes from the MCU, at any baud,
-  in any reset/download mode (see [lessons.md](../../work/past/lessons.md), which cost an hour of chasing a
-  cable that wasn't the problem).
+- **Buttons**: BOOT (GPIO61), RESET (EN).
+- **J5, power/enable jumper** (a 2-pin header next to the 5V→3.3V DC/DC converter U4 and the 3.3V
+  power-on LED D11). **Leave it installed.** With J5 removed the board is *half-powered*: the CP2102N runs off USB VBUS so its port still enumerates, but the ESP32-S31's 3.3V/EN rail is incomplete and the chip drives nothing, you get a serial port that opens but zero bytes from the MCU, at any baud, in any reset/download mode (see [lessons.md](../../work/past/lessons.md), which cost an hour of chasing a cable that wasn't the problem).
 - **40-pin GPIO header** (J2). Optional 32.768 kHz crystal footprint (Y1, NC by default).
 
 ## Free GPIO for user peripherals (LED strips, loopback)
@@ -109,32 +96,26 @@ The board's own peripherals claim a large, contiguous low-GPIO block; the **J2 h
 | Onboard WS2812 RGB LED | 60 |
 | BOOT button | 61 |
 
-**J2 is a 2×20 header** — each column carries two GPIOs, a top-row pin above a bottom-row pin. Reading the silkscreen left→right (top / bottom per column):
+**J2 is a 2×20 header**, each column carries two GPIOs, a top-row pin above a bottom-row pin. Reading the silkscreen left→right (top / bottom per column):
 
 | col | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 |
 |---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
 | **top** | G | RX0 | 60 | G | 3 | 1 | **48** | 46 | 44 | 42 | 38 | 36 | D0 | D2 | CLK | 4 | C | 3V3 | G | 5V |
 | **bottom** | G | TX0 | 61 | 2 | 0 | 49 | **47** | 45 | 43 | 40 | 39 | 37 | 35 | D1 | D3 | CMD | G | 3V3 | G | 5V |
 
-The two pins in **one column are physically stacked**, so a 2-pin jumper cap bridges them with no flying wire — that adjacency is what makes a column a good loopback pair.
+The two pins in **one column are physically stacked**, so a 2-pin jumper cap bridges them with no flying wire, that adjacency is what makes a column a good loopback pair.
 
 **Free on J2 for user I/O** (read off the table above, minus the board peripherals and boot straps): the numbered GPIOs on cols 5–16 (**4, 36, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 48, 49**) are plain I/O clear of Ethernet (2, 5–19), audio (50–57), the SD lines (broken out as `D0`–`D3` / `CLK` / `CMD` by function, cols 13–16), the onboard LED (60) and the straps (0, 1, 3, 61). The `C` label at col 17 is a chip-enable, not a GPIO. **The GPIO numbers here are read from the board silkscreen and not yet bench-confirmed**: the S31 reference pin tables have been found off-by-one before (see the [S31 Ethernet lesson](../../work/past/lessons.md)), so probe a pin before committing a design to it.
 
 **Recommended assignment** (what the S31 catalog entry uses):
 
-- **LED strip data:** the onboard WS2812 is on **GPIO 60** (the catalog default). For an *external* strand, use **GPIO 42** as the single-lane pick; a parallel rig (RMT/Parlio) takes the free block (**36–49**, skipping 41 which isn't broken out) for several lanes. **GPIO 4** (col 16 top) also works as an LED data pin and sits one column from the `G` / `3V3` / `5V` power rail (cols 17–20), so a single strip's data + ground + 5 V wires land close together — handy for a tidy 3-wire pigtail. It's a plain I/O with no strap or peripheral tie on this board (the SD lines beside it, D0–D3 / CLK / CMD, are broken out by function name, not GPIO number, so GPIO 4 is *not* one of them; it just neighbors that cluster on the header). The only reason it reads as "distinct" from the rest of the free run is its header position — it's over by the SD/power group rather than in the low-block on cols 8–12.
-- **Loopback self-test:** **Tx = GPIO 48, Rx = GPIO 47** — the two pins of **column 7** (48 top, 47 bottom), so a single jumper cap shorts them. A driver transmits a known WS2812 frame out Tx and reads it back on Rx to verify output on real silicon (same pattern as the P4-NANO bench's 32↔33). They sit at the top of the free run, clear of the operational LED pins so the strip wiring and the jumper don't interfere. **Bench-confirmed on the S31 for [RMT](../../moonmodules/light/drivers.md#rmtled) and the [Parallel LED driver](../../moonmodules/light/drivers.md#parallelled) across all three of its peripherals.** The S31 SOC has a real LCD_CAM (`SOC_LCDCAM_I80_LCD_SUPPORTED`), so its `i80` backend is LCD_CAM-driven (not the classic ESP32's I2S-in-i80-mode, which the platform layer excludes on any chip with real LCD_CAM) and can draw its frame from PSRAM. The `peripheral` selector therefore offers all three: **`i80`, `MoonI80`, and `Parlio`** — the RGMII Ethernet does not take the LCD_CAM block. Bench-verified: an 8x8 panel driving on `i80` (data GPIO 60), and Parlio on the free-block pins. Testing several drivers in a row, they all default loopback to GPIO 48, so only one can hold the pin at a time — toggle each driver's `loopbackTest` off before testing the next.
+- **LED strip data:** the onboard WS2812 is on **GPIO 60** (the catalog default). For an *external* strand, use **GPIO 42** as the single-lane pick; a parallel rig (RMT/Parlio) takes the free block (**36–49**, skipping 41 which isn't broken out) for several lanes. **GPIO 4** (col 16 top) also works as an LED data pin and sits one column from the `G` / `3V3` / `5V` power rail (cols 17–20), so a single strip's data + ground + 5 V wires land close together, handy for a tidy 3-wire pigtail. It's a plain I/O with no strap or peripheral tie on this board (the SD lines beside it, D0–D3 / CLK / CMD, are broken out by function name, not GPIO number, so GPIO 4 is *not* one of them; it just neighbors that cluster on the header). The only reason it reads as "distinct" from the rest of the free run is its header position, it's over by the SD/power group rather than in the low-block on cols 8–12.
+- **Loopback self-test:** **Tx = GPIO 48, Rx = GPIO 47**, the two pins of **column 7** (48 top, 47 bottom), so a single jumper cap shorts them. A driver transmits a known WS2812 frame out Tx and reads it back on Rx to verify output on real silicon (same pattern as the P4-NANO bench's 32↔33). They sit at the top of the free run, clear of the operational LED pins so the strip wiring and the jumper don't interfere. **Bench-confirmed on the S31 for [RMT](../../moonmodules/light/drivers.md#rmtled) and the [Parallel LED driver](../../moonmodules/light/drivers.md#parallelled) across all three of its peripherals.** The S31 SOC has a real LCD_CAM (`SOC_LCDCAM_I80_LCD_SUPPORTED`), so its `i80` backend is LCD_CAM-driven (not the classic ESP32's I2S-in-i80-mode, which the platform layer excludes on any chip with real LCD_CAM) and can draw its frame from PSRAM. The `peripheral` selector therefore offers all three: **`i80`, `MoonI80`, and `Parlio`**, the RGMII Ethernet does not take the LCD_CAM block. Bench-verified: an 8x8 panel driving on `i80` (data GPIO 60), and Parlio on the free-block pins. Testing several drivers in a row, they all default loopback to GPIO 48, so only one can hold the pin at a time, toggle each driver's `loopbackTest` off before testing the next.
 
 ## SoC capabilities (from `components/soc/esp32s31/include/soc/soc_caps.h`)
 
-Wi-Fi 6 · Bluetooth (no separate BLE soc-flag) · IEEE 802.15.4 (Thread/Zigbee) · USB-OTG · GPSPI ·
-TWAI (CAN) · RMT · Parlio · LCD_CAM i80 · on-chip EMAC · PSRAM. RISC-V dual-core.
+Wi-Fi 6 · Bluetooth (no separate BLE soc-flag) · IEEE 802.15.4 (Thread/Zigbee) · USB-OTG · GPSPI · TWAI (CAN) · RMT · Parlio · LCD_CAM i80 · on-chip EMAC · PSRAM. RISC-V dual-core.
 
 The S31 catalog entry drives **LEDs** (RMT on GPIO60) and **Wi-Fi 6**, and wires an
 **[I2cScanModule](../../moonmodules/core/moxygen/I2cScanModule.md)** on the codec bus (SDA 51 / SCL 50) for
-I2C bring-up. The **[AudioService](../../moonmodules/core/moxygen/AudioService.md)** ES8311 path is implemented
-— the codec seam configures the ES8311 over I2C (codec reachable, ACK at 0x18) and AudioService
-reads the I2S mic. End-to-end mic validation depends on confirming MCLK at GPIO52; the S31 entry
-keeps **Audio** under `planned` until that bench check passes, so the installer advertises only
-what's confirmed working. The other board capabilities (Ethernet, Bluetooth, SD, USB host, …)
-likewise live in the entry's `planned` list — see the S31 entry in `mooninstaller/deviceModels.json`.
+I2C bring-up. The **[AudioService](../../moonmodules/core/moxygen/AudioService.md)** ES8311 path is implemented, the codec seam configures the ES8311 over I2C (codec reachable, ACK at 0x18) and AudioService reads the I2S mic. End-to-end mic validation depends on confirming MCLK at GPIO52; the S31 entry keeps **Audio** under `planned` until that bench check passes, so the installer advertises only what's confirmed working. The other board capabilities (Ethernet, Bluetooth, SD, USB host, …) likewise live in the entry's `planned` list, see the S31 entry in `mooninstaller/deviceModels.json`.
diff --git a/docs/reference/hardware/light-fixtures.md b/docs/reference/hardware/light-fixtures.md
index 4f4dd20f..c4bb5804 100644
--- a/docs/reference/hardware/light-fixtures.md
+++ b/docs/reference/hardware/light-fixtures.md
@@ -1,25 +1,18 @@
 # Light fixtures and DMX nodes: hardware reference
 
-DMX channel maps and specifications for the fixtures and Art-Net nodes on the bench, read from
-their manuals so a light preset can be built without re-scraping a PDF. A fixture here has a
-matching entry in the **LightPresets** library ([drivers](../../moonmodules/light/drivers.md)); the
-preset is the channel-role layout, and the numbers behind it are below.
+DMX channel maps and specifications for the fixtures and Art-Net nodes on the bench, read from their manuals so a light preset can be built without re-scraping a PDF. A fixture here has a matching entry in the **LightPresets** library ([drivers](../../moonmodules/light/drivers.md)); the preset is the channel-role layout, and the numbers behind it are below.
 
 ## SHEHDS Mini LED Moving Head 10W RGBW
 
 ![SHEHDS Mini LED moving head](../../assets/light/fixtures/djlight-mini-led-10w.jpg){ width="280" }
 
-A single 10 W RGBW 4-in-1 LED in a compact moving head: the smallest fixture that exercises the
-full pan/tilt/color path, which is why it is the first moving head on the bench.
+A single 10 W RGBW 4-in-1 LED in a compact moving head: the smallest fixture that exercises the full pan/tilt/color path, which is why it is the first moving head on the bench.
 
 **Source:** the printed manual supplied with the fixture
-([channel tables](../../assets/light/fixtures/djlight-mini-led-10w-dmx-channels.jpg),
-[display menu](../../assets/light/fixtures/djlight-mini-led-10w-display.jpg)) ·
-[product listing](https://www.amazon.com/SHEHDS-Lighting-Professional-Channels-Christmas/dp/B08B4MV36K)
+([channel tables](../../assets/light/fixtures/djlight-mini-led-10w-dmx-channels.jpg), [display menu](../../assets/light/fixtures/djlight-mini-led-10w-display.jpg)) · [product listing](https://www.amazon.com/SHEHDS-Lighting-Professional-Channels-Christmas/dp/B08B4MV36K)
 
 A PDF circulating under this ASIN describes a 9/14-channel fixture with a combined shutter/dimmer.
-That is a DIFFERENT model: this one is 11/13 channels with a plain linear dimmer and a separate
-strobe channel. Trust the printed manual; the tables below come from it.
+That is a DIFFERENT model: this one is 11/13 channels with a plain linear dimmer and a separate strobe channel. Trust the printed manual; the tables below come from it.
 
 | | |
 |---|---|
@@ -47,28 +40,19 @@ strobe channel. Trust the printed manual; the tables below come from it.
 | 10 | 0-255 | blue, 0 = closed | `B` |
 | 11 | 0-255 | white, 0 = closed | `W` |
 
-CH6 is a straight linear dimmer. Today the driver holds it fully open and keeps brightness in the
-color values; routing brightness onto it is the better model and is
-[backlogged](../../work/future/backlog-light.md). The channels left
-unmapped hold at 0, which is what a light driver wants: **strobe off** (CH7) and full-speed
-movement (CH5). The fine channels are unused until 16-bit positioning is wired up.
+CH6 is a straight linear dimmer. Today the driver holds it fully open and keeps brightness in the color values; routing brightness onto it is the better model and is [backlogged](../../work/future/backlog-light.md). The channels left unmapped hold at 0, which is what a light driver wants: **strobe off** (CH7) and full-speed movement (CH5). The fine channels are unused until 16-bit positioning is wired up.
 
-Two things matter when driving it by hand. Every channel at 255 makes the fixture **strobe**, because
-CH7 is a strobe channel rather than part of the dimmer. And the 13-channel mode's CH12 at 250-255
-selects **sound mode**, where the fixture ignores DMX movement and runs its own program.
+Two things matter when driving it by hand. Every channel at 255 makes the fixture **strobe**, because CH7 is a strobe channel rather than part of the dimmer. And the 13-channel mode's CH12 at 250-255 selects **sound mode**, where the fixture ignores DMX movement and runs its own program.
 
 ### 13-channel mode (the alternative)
 
-The same first 11 channels, plus CH12 (0-249 auto run, 250-255 sound mode) and CH13 (150-200
-reset). Both are program modes that take control away from DMX, so the 11-channel mode is the one
-worth driving from projectMM.
+The same first 11 channels, plus CH12 (0-249 auto run, 250-255 sound mode) and CH13 (150-200 reset). Both are program modes that take control away from DMX, so the 11-channel mode is the one worth driving from projectMM.
 
 ## P-Knight Art-Net2 CR021R
 
 ![P-Knight Art-Net2 CR021R](../../assets/light/fixtures/pknight-artnet2-cr021r.jpg){ width="280" }
 
-A two-universe Art-Net to DMX512 node: Ethernet in, XLR DMX out. It is how projectMM drives a wired
-fixture, the counterpart to the LED drivers that speak to addressable strips directly.
+A two-universe Art-Net to DMX512 node: Ethernet in, XLR DMX out. It is how projectMM drives a wired fixture, the counterpart to the LED drivers that speak to addressable strips directly.
 
 | | |
 |---|---|
@@ -79,16 +63,9 @@ fixture, the counterpart to the LED drivers that speak to addressable strips dir
 | Bench universe | `01` on its display = `universe_start` **1** in NetworkSendDriver |
 
 **Its universe display is 1-based: node `01` is Art-Net universe 1.** Set the driver's
-`universe_start` to the number the node shows, not one less. This is the setting that cost a whole
-bench session. A universe mismatch is silent: ArtDMX for another universe is dropped with no error
-anywhere, and the symptom is identical to a broken fixture, a wrong DMX
-address or a closed shutter. Check the universe FIRST when a fixture does not respond, and confirm
-it by reading the number off the node's own display.
+`universe_start` to the number the node shows, not one less. This is the setting that cost a whole bench session. A universe mismatch is silent: ArtDMX for another universe is dropped with no error anywhere, and the symptom is identical to a broken fixture, a wrong DMX address or a closed shutter. Check the universe FIRST when a fixture does not respond, and confirm it by reading the number off the node's own display.
 
-Note the fixture's own DMX address is 1-based too but starts at slot 1: address `a0001` means its
-CH1 is DMX slot 1, so an 11-channel fixture at address 1 occupies slots 1-11.
+Note the fixture's own DMX address is 1-based too but starts at slot 1: address `a0001` means its CH1 is DMX slot 1, so an 11-channel fixture at address 1 occupies slots 1-11.
 
 **It answers no TCP and no ArtPoll.** Verified on the bench: ports 80, 6454, 8080 and 23 are all
-closed to TCP, and an ArtPoll broadcast draws no ArtPollReply. The node is a pure receiver, so it
-does not appear in a discovery scan and its address is set from its own display, not over the
-network. Send ArtDMX to its address and it outputs; there is nothing to query.
+closed to TCP, and an ArtPoll broadcast draws no ArtPollReply. The node is a pure receiver, so it does not appear in a discovery scan and its address is set from its own display, not over the network. Send ArtDMX to its address and it outputs; there is nothing to query.
diff --git a/docs/reference/metrics/repo-health.json b/docs/reference/metrics/repo-health.json
index b47355f0..36bd0673 100644
--- a/docs/reference/metrics/repo-health.json
+++ b/docs/reference/metrics/repo-health.json
@@ -1,13 +1,13 @@
 {
-  "commit": "7bfbfa78",
+  "commit": "f1be21c4",
   "flash": {
     "esp32s3-n16r8": 2134032,
-    "desktop": 1970568,
-    "esp32": 2080464,
-    "esp32p4rev1-eth": 2029776,
+    "desktop": 1970664,
+    "esp32": 2084672,
+    "esp32p4rev1-eth": 2034288,
     "esp32p4rev1-eth-wifi": 2284640,
     "esp32s3-n8r8": 2087168,
-    "esp32s31": 2348592,
+    "esp32s31": 2427440,
     "esp32-16mb": 2060368,
     "esp32-eth": 1681040,
     "esp32-wrover": 1843760,
@@ -17,9 +17,9 @@
     "esp32-pico": 2107168
   },
   "measured": {
-    "esp32p4rev1-eth": "2026-09-18",
-    "esp32s31": "2026-09-06",
-    "esp32": "2026-09-20",
+    "esp32p4rev1-eth": "2026-09-21",
+    "esp32s31": "2026-09-21",
+    "esp32": "2026-09-21",
     "esp32-pico": "2026-09-09",
     "esp32s3-n16r8": "2026-09-21",
     "desktop": "2026-09-21",
@@ -31,8 +31,8 @@
   },
   "perf": {
     "desktop": {
-      "tick_us": 135,
-      "fps": 7407,
+      "tick_us": 124,
+      "fps": 8064,
       "scenario_p50": {
         "Layer_base_pipeline": {
           "p50": 70,
@@ -698,54 +698,54 @@
     }
   },
   "loc": {
-    "core": 21994,
-    "light": 29903,
+    "core": 22010,
+    "light": 29952,
     "platform": 16346,
-    "ui": 11223,
-    "test": 56265,
-    "moondeck": 27113
+    "ui": 11310,
+    "test": 56377,
+    "moondeck": 27235
   },
   "comments": {
     "core": {
-      "lines": 5575,
+      "lines": 5579,
       "ratio": 0.28
     },
     "light": {
-      "lines": 7262,
-      "ratio": 0.273
+      "lines": 7269,
+      "ratio": 0.272
     },
     "platform": {
       "lines": 3481,
       "ratio": 0.238
     },
     "ui": {
-      "lines": 3350,
+      "lines": 3371,
       "ratio": 0.315
     },
     "test": {
-      "lines": 6345,
+      "lines": 6360,
       "ratio": 0.131
     },
     "moondeck": {
-      "lines": 4531,
+      "lines": 4570,
       "ratio": 0.191
     }
   },
   "tests": {
-    "cases": 2067,
+    "cases": 2070,
     "scenarios": 27
   },
   "docs": {
     "md_files": 138,
-    "md_lines": 28986,
+    "md_lines": 28265,
     "plans_files": 31,
     "backlog_lines": 3025,
     "lessons_lines": 518,
     "claude_md_lines": 279
   },
   "complexity": {
-    "functions": 3627,
-    "over_threshold": 265,
+    "functions": 3630,
+    "over_threshold": 267,
     "worst_ccn": 128
   }
 }
diff --git a/docs/reference/metrics/repo-health.md b/docs/reference/metrics/repo-health.md
index 851eb4f3..5178a840 100644
--- a/docs/reference/metrics/repo-health.md
+++ b/docs/reference/metrics/repo-health.md
@@ -1,6 +1,6 @@
 # Repo health
 
-Measured at `7bfbfa78`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
+Measured at `f1be21c4`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
 
 Current state only; the trend is this file's git history (`git log -p docs/reference/metrics/repo-health.md`). Nothing here fails a build: the numbers make growth visible, the judgment stays human.
 
@@ -8,19 +8,19 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Flash | Capacity | Used | Built |
 |---|---:|---:|---:|:--:|
-| desktop | 1,924 KB (+17 KB) ⚠ | - | - | yes |
-| esp32 | 2,032 KB | 2,496 KB | 81% | carried 1d |
+| desktop | 1,924 KB (+96 B) ⚠ | - | - | yes |
+| esp32 | 2,036 KB (+4 KB) ⚠ | 2,496 KB | 82% | yes |
 | esp32-16mb | 2,012 KB | 4,096 KB | 49% | **STALE 12d** |
 | esp32-eth | 1,642 KB | 2,496 KB | 66% | **STALE 10d** |
 | esp32-pico | 2,058 KB | 3,072 KB | 67% | **STALE 12d** |
 | esp32-wrover | 1,801 KB | - | - | carried (age?) |
-| esp32p4rev1-eth | 1,982 KB | 4,096 KB | 48% | carried 3d |
+| esp32p4rev1-eth | 1,987 KB (+4 KB) ⚠ | 4,096 KB | 49% | yes |
 | esp32p4rev1-eth-wifi | 2,231 KB | 4,096 KB | 54% | **STALE 13d** |
 | esp32p4rev3-eth | 1,605 KB | - | - | carried (age?) |
-| esp32s3-n16r8 | 2,084 KB (+2 KB) ⚠ | 4,096 KB | 51% | yes |
+| esp32s3-n16r8 | 2,084 KB | 4,096 KB | 51% | yes |
 | esp32s3-n8r8 | 2,038 KB | 3,072 KB | 66% | **STALE 13d** |
 | esp32s3-zero | 1,977 KB | 2,496 KB | 79% | **STALE 13d** |
-| esp32s31 | 2,294 KB | 4,096 KB | 56% | **STALE 15d** |
+| esp32s31 | 2,371 KB (+77 KB) ⚠ | 4,096 KB | 58% | yes |
 | qemu | 1,351 KB | - | - | carried (age?) |
 
 `Built: yes` was measured this run. `carried (age?)` was not rebuilt either and predates this record, so its age is unknown: it dates itself on the next build. `carried Nd` was NOT rebuilt and its number is N days old, so an absent delta says nothing about the change. **STALE** marks a carry older than 7 days: the number has gone unchecked long enough that growth will surface later as one jump, blamed on whichever commit happens to rebuild that target. `Used` is against the app slot in the firmware's own partition table.
@@ -29,7 +29,7 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Tick | FPS |
 |---|---:|---:|
-| desktop | 135 µs (+15 µs) ⚠ | 7,407 (−926) ⚠ |
+| desktop | 124 µs (−11 µs) ✓ | 8,064 (+657) ✓ |
 | esp32 | 8,354 µs | 119 |
 
 ### Scenario tick by target (p50 of each sample window)
@@ -37,10 +37,10 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 | Scenario | desktop-macos | desktop-windows | esp32 | esp32s3-n16r8 | esp32p4rev1-eth | esp32s31 | esp32-eth | esp32-eth-wifi | unknown |
 |---|---|---|---|---|---|---|---|---|---|
 | Audio_mutation | 22 | 40 ? | 13,152 | 47 ? | - | - | - | - | - |
-| Aurora_fps | 1,521 (−5) ✓ | - | - | - | - | - | - | - | - |
+| Aurora_fps | 1,521 | - | - | - | - | - | - | - | - |
 | Driver_mutation | 20 | 42 ? | 12,812 | 39 ? | - | - | - | - | - |
 | Effects_composition | 144 | 549 ? | - | - | - | - | - | - | - |
-| Fields_polar_lut | 1,263 (−12) ✓ | - | - | - | - | - | - | - | - |
+| Fields_polar_lut | 1,263 | - | - | - | - | - | - | - | - |
 | Fluid_solver | 217 | - | - | - | - | - | - | - | - |
 | GridBlacks_blackpixel | 2 | 8 ? | 269 ? | 267 ? | - | - | - | - | - |
 | GridLayout_resize | 120 | 219 ? | 1,352 ? | 1,011 ? | 1,143 ? | - | 95,771 ? | 82,231 ? | - |
@@ -56,8 +56,8 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 | MultiplyModifier_pipeline | 119 | 225 ? | - | - | - | - | - | - | - |
 | NetworkModule_eth_reconfigure | - | - | 1,169 ? | 97,843 ? | - | - | - | - | - |
 | NetworkModule_mdns_toggle | 13 ? | - | 36 ? | 36 ? | 21 ? | - | 109,767 ? | 93,963 ? | - |
-| Trails_ladder | 358 (−1) ✓ | - | - | - | - | - | - | - | - |
-| modifier_chain | 43 (−1) ✓ | 69 ? | 13,337 | - | - | - | - | - | - |
+| Trails_ladder | 358 | - | - | - | - | - | - | - | - |
+| modifier_chain | 43 | 69 ? | 13,337 | - | - | - | - | - | - |
 | modifier_swap | 23 | 41 ? | 12,250 | 354 ? | 362 ? | - | 1,010 ? | - | - |
 | perf_full | 251 | 592 ? | 10,392 | 16,915 ? | 17,433 ? | - | - | - | - |
 | perf_light | 15 | 35 ? | 2,183 | 2,485 ? | 2,038 ? | - | - | - | - |
@@ -81,36 +81,36 @@ These build a bare pipeline with no optional modules, so a change here is a chan
 
 | Area | Lines | Comments | Comment share |
 |---|---:|---:|---:|
-| core | 21,994 (+145) ⚠ | 5,575 | 28.0 % |
-| light | 29,903 (+16) ⚠ | 7,262 | 27.3 % (+0.1 %) ⚠ |
-| platform | 16,346 (+21) ⚠ | 3,481 | 23.8 % |
-| ui | 11,223 (+27) ⚠ | 3,350 | 31.5 % (+0.1 %) ⚠ |
-| test | 56,265 (+108) ⚠ | 6,345 | 13.1 % |
-| moondeck | 27,113 (+193) ⚠ | 4,531 | 19.1 % (+0.1 %) ⚠ |
+| core | 22,010 (+16) ⚠ | 5,579 | 28.0 % |
+| light | 29,952 (+49) ⚠ | 7,269 | 27.2 % (−0.1 %) ✓ |
+| platform | 16,346 | 3,481 | 23.8 % |
+| ui | 11,310 (+87) ⚠ | 3,371 | 31.5 % |
+| test | 56,377 (+112) ⚠ | 6,360 | 13.1 % |
+| moondeck | 27,235 (+122) ⚠ | 4,570 | 19.1 % |
 
 ## Tests
 
 | Kind | Count |
 |---|---:|
-| unit cases | 2,067 (+2) ✓ |
+| unit cases | 2,070 (+3) ✓ |
 | scenarios | 27 |
 
 ## Complexity
 
 | Metric | Value |
 |---|---:|
-| functions | 3,627 (+6) ✓ |
-| over threshold | 265 (+1) ⚠ |
+| functions | 3,630 (+3) ✓ |
+| over threshold | 267 (+2) ⚠ |
 | worst CCN | 128 |
 
 ## Documentation
 
 | Metric | Value |
 |---|---:|
-| markdown files | 138 (+1) ⚠ |
-| markdown lines | 28,986 (+93) ⚠ |
+| markdown files | 138 |
+| markdown lines | 28,265 (−721) ✓ |
 | plan files | 31 |
-| backlog lines | 3,025 (+39) ⚠ |
+| backlog lines | 3,025 |
 | lessons lines | 518 |
-| CLAUDE.md lines | 279 (+2) ⚠ |
+| CLAUDE.md lines | 279 |
 
diff --git a/docs/reference/performance.md b/docs/reference/performance.md
index d70f1cfb..f5c6f03f 100644
--- a/docs/reference/performance.md
+++ b/docs/reference/performance.md
@@ -64,9 +64,7 @@ Binary sizes:
 | Tunnel (fbm8) | 16,385 | 21,304 | 16,649 | 1.02 |
 | PolarNoise (warp8) | 20,356 | 29,199 | 20,490 | 1.01 |
 
-The two noise effects have since merged: `Noise` is `Dim::D3` and renders what `Noise2D` did, so the
-two rows above are one effect's 2D and 3D paths under the names they carried when the swap was
-measured.
+The two noise effects have since merged: `Noise` is `Dim::D3` and renders what `Noise2D` did, so the two rows above are one effect's 2D and 3D paths under the names they carried when the swap was measured.
 
 The 3D path is the closest to the bound and the reason: 3D gradient noise does eight dot products the value form never did, and the S3 instruction count for a 3D sample is 1.3x the old one. Method worth keeping: compile the kernel with the target's own compiler (`xtensa-esp32s3-elf-g++ -O2 -S`) and count instructions, branches and stack spills BEFORE flashing; three restructurings were compared that way in seconds, and the one flash went to the winner. The P4 and S31 numbers are open until those boards are back on the bench.
 
@@ -95,9 +93,7 @@ That reading held: the 8-bit tier improved most (2D by 4x on the host), because
 
 ### Fluid solver cost (host)
 
-`scenario_Fluid_solver`, desktop macOS arm64, tick in µs. The solver is Stam's stable fluid: several
-passes over the grid per frame, plus `iterations` Gauss-Seidel sweeps for the pressure projection
-that keeps the flow divergence-free.
+`scenario_Fluid_solver`, desktop macOS arm64, tick in µs. The solver is Stam's stable fluid: several passes over the grid per frame, plus `iterations` Gauss-Seidel sweeps for the pressure projection that keeps the flow divergence-free.
 
 | Grid | iterations | tick µs |
 |---|---:|---:|
@@ -108,16 +104,11 @@ that keeps the flow divergence-free.
 | 20×20×20 cube | 5 | 249 |
 | 16×16 | 5 | 7 |
 
-Two properties an author picks a setting from. **`iterations` is near-linear**: 1 to 20 is 20 to
-69 µs, since each is another sweep over the whole grid. **The forcing is free next to the solver**:
+Two properties an author picks a setting from. **`iterations` is near-linear**: 1 to 20 is 20 to 69 µs, since each is another sweep over the whole grid. **The forcing is free next to the solver**:
 going from 2 jets to 4, and persistence from 150 to 255, moved 135 µs to 134 µs, inside the noise.
-So the grid and the iteration count are the two knobs that matter, and the jets are a look rather
-than a cost. **A cube is depth times one panel**: twenty 20×20 slices, each its own medium, cost
-249 µs against 133 for one 64×64 panel with about half the lights, which is the per-slice solve
-paying its boundary and projection twenty times over.
+So the grid and the iteration count are the two knobs that matter, and the jets are a look rather than a cost. **A cube is depth times one panel**: twenty 20×20 slices, each its own medium, cost 249 µs against 133 for one 64×64 panel with about half the lights, which is the per-slice solve paying its boundary and projection twenty times over.
 
-Device rows are open: the P4 and S3 numbers need a board and have not been measured, so what
-this effect can carry on either is an open question rather than a claim.
+Device rows are open: the P4 and S3 numbers need a board and have not been measured, so what this effect can carry on either is an open question rather than a claim.
 
 ### Memory at 128×128 with mirror
 
@@ -261,8 +252,7 @@ The exec block is the emitted machine code, so it varies by ISA (the RISC-V rows
 The **depth guard** is one arena byte, incremented on entry and decremented in the epilogue. A refused call returns rather than the caller branching around it, which is why the cost sits in the callee and not at every call site. Unbounded recursion therefore degrades instead of resetting: the classic ran a deliberately non-terminating script for 110 s at 109 fps, with the deepest calls doing nothing.
 
 **Two cost models** (2026-08-22, shiffy's 80x48 = 3,840 lights). A shader is a function of position
-and time, so it pays per LIGHT; a particle script pays per OBJECT, with the per-particle work inside
-one C++ loop per call. This is the first script vocabulary where that distinction shows.
+and time, so it pays per LIGHT; a particle script pays per OBJECT, with the per-particle work inside one C++ loop per call. This is the first script vocabulary where that distinction shows.
 
 | Script | Shape | shiffy 80x48 | desktop 128x96 |
 |---|---|---:|---:|
@@ -271,35 +261,18 @@ one C++ loop per call. This is the first script vocabulary where that distinctio
 | `fountain.mle` | ~9 per FRAME, 300 particles | 1,093 us | 9 us |
 | `ballpit.mle` | as above plus `collide` over 64 | 5,127 us | |
 
-`metal.mle` against `fountain.mle` is 54x on the same fixture. `polarR` is what makes the shader
-expensive: it wraps a real square root, measured at ~3.5 us per pixel for that one builtin, and
-`metal` calls it three times per pixel. `ballpit` shows `collide`'s N-body cost, which is the one
-call here that is not linear in pool size: 3.2 us at 48 particles against 0.1 us without, 53.6 us at
-200 (host figures; an S3 is 20-40x slower).
+`metal.mle` against `fountain.mle` is 54x on the same fixture. `polarR` is what makes the shader expensive: it wraps a real square root, measured at ~3.5 us per pixel for that one builtin, and `metal` calls it three times per pixel. `ballpit` shows `collide`'s N-body cost, which is the one call here that is not linear in pool size: 3.2 us at 48 particles against 0.1 us without, 53.6 us at 200 (host figures; an S3 is 20-40x slower).
 
 **A 1 Hz filesystem scan was stuttering every device.** `FileManagerModule::tick1s()` called
-`esp_littlefs_info`, which walks every block of the partition (~80 ms on an S3), inline on the render
-thread, to feed one progress bar. Frame deltas per second on shiffy went from
-`83 78 80 79 82 66 72` to `83 85 85 83 84 87 85` once it was throttled to once a minute: the dip is
-gone and average throughput rose from ~77 to ~85 fps. It is pre-existing, and particles are what made
-it visible, because a particle INTEGRATES a stall into its trajectory where a shader redraws past it.
+`esp_littlefs_info`, which walks every block of the partition (~80 ms on an S3), inline on the render thread, to feed one progress bar. Frame deltas per second on shiffy went from `83 78 80 79 82 66 72` to `83 85 85 83 84 87 85` once it was throttled to once a minute: the dip is gone and average throughput rose from ~77 to ~85 fps. It is pre-existing, and particles are what made it visible, because a particle INTEGRATES a stall into its trajectory where a shader redraws past it.
 One frame after an 80 ms gap moves every particle 6.7x its usual distance.
 
 **Desktop tick across this cycle:** 150 → 133 µs (6666 → 7518 fps), measured by `collect_kpi.py --commit` at each commit. The gain is not from MoonLive — it tracks the two heap-overrun fixes and the register-reuse work landing earlier in the branch. No scenario `contract` was renegotiated on this branch: all 20 scenarios pass inside their existing budgets, which is the assertion surface this page defers to.
 
 **The compile-time staging buffer is sized from the script's tokens**, at 48 bytes per token plus a
-256-byte floor, and freed when the compile returns. The constant is the worst case rather than the
-average, because the buffer is allocated before a byte is emitted: measured across every shipped
-script on all three backends, the densest is `random-pixel.mlv` at 28.5 B/token on RISC-V (one
-statement, four nested `random16()` calls, each saving the whole register pool), so 48 is a ~1.7x
-margin. Density FALLS as a script grows, so a short call-dense script sets the bound: `gradient.mlv`
-is 5.9 and the longest shipped script is 15.3.
-
-It was 64, measured before host arguments moved into frame slots shrank what a call saves. At that
-figure the two longest scripts asked for more than the 16 KB sanity bound and were served by the
-clamp, so a script's buffer had stopped tracking its size. Re-measuring took 25% off the transient
-allocation, which matters on a classic ESP32 where the compile shares a 12 KB task. A per-ISA test
-pins that every shipped script still emits under two-thirds of its budget.
+256-byte floor, and freed when the compile returns. The constant is the worst case rather than the average, because the buffer is allocated before a byte is emitted: measured across every shipped script on all three backends, the densest is `random-pixel.mlv` at 28.5 B/token on RISC-V (one statement, four nested `random16()` calls, each saving the whole register pool), so 48 is a ~1.7x margin. Density FALLS as a script grows, so a short call-dense script sets the bound: `gradient.mlv` is 5.9 and the longest shipped script is 15.3.
+
+It was 64, measured before host arguments moved into frame slots shrank what a call saves. At that figure the two longest scripts asked for more than the 16 KB sanity bound and were served by the clamp, so a script's buffer had stopped tracking its size. Re-measuring took 25% off the transient allocation, which matters on a classic ESP32 where the compile shares a 12 KB task. A per-ISA test pins that every shipped script still emits under two-thirds of its budget.
 
 **Flash**, measured by building the classic at the branch point and again with the local-call work: 1723295 → 1726027 bytes, +2732 (+0.16%). High per line of source (about 20 bytes for ~137 net lines of code) because nearly all of it is emitter code instantiated once per backend, so one line of the shared lowering becomes three copies of emitted-instruction sequences in the image.
 
@@ -332,8 +305,7 @@ The **acceptance floors** these establish for the parallel backends: RMT **8×25
 
 ## Panel cards over raw Ethernet (`PanelCardDriver`, ESP32-S31)
 
-Measured on an S31 driving two 128x64 HUB75 panels through a ColorLight 5A-75 receiver card over a
-gigabit RGMII link, 2026-07-31.
+Measured on an S31 driving two 128x64 HUB75 panels through a ColorLight 5A-75 receiver card over a gigabit RGMII link, 2026-07-31.
 
 | | us/tick | note |
 |---|---:|---|
@@ -341,31 +313,20 @@ gigabit RGMII link, 2026-07-31.
 | PreviewDriver | 5 687 | the browser preview, same buffer |
 | a heavy effect (GameOfLife) | 17 592 | the render, and the largest single cost |
 
-The panel driver is the cheapest active module despite pushing 132 packets per frame (2 brightness +
-128 rows + 2 sync, at 497 pixels per row packet). The wall's frame rate is set by the render, not by
-the output: a heavy effect at 17.6 ms dominates a 26 ms tick, giving ~32 FPS, while a lighter effect
-mix measures ~56 FPS on the same wall.
+The panel driver is the cheapest active module despite pushing 132 packets per frame (2 brightness + 128 rows + 2 sync, at 497 pixels per row packet). The wall's frame rate is set by the render, not by the output: a heavy effect at 17.6 ms dominates a 26 ms tick, giving ~32 FPS, while a lighter effect mix measures ~56 FPS on the same wall.
 
 **The ceiling is packets, not pixels.** Each frame is sent synchronously from `tick()`, so a taller
-wall costs proportionally more rows; a 256-row wall would double the packet count. The card format's
-1 Gbit requirement is a wire-time constraint rather than a bandwidth one: at 100 Mbit the same bytes
-take ten times as long and overrun the inter-frame window the sync depends on
-([drivers.md](../moonmodules/light/drivers.md#panelcard)).
+wall costs proportionally more rows; a 256-row wall would double the packet count. The card format's 1 Gbit requirement is a wire-time constraint rather than a bandwidth one: at 100 Mbit the same bytes take ten times as long and overrun the inter-frame window the sync depends on ([drivers.md](../moonmodules/light/drivers.md#panelcard)).
 
 **The DMA ring is what makes it stable.** `CONFIG_ETH_DMA_BUFFER_SIZE` defaults to 512 B, so a
-1512 B frame spanned three descriptors and a 10-descriptor ring held ~3.3 frames while the driver
-fires 132 back-to-back. At that depth the S31 refused ~19 000 frames and wedged twice inside 20
-minutes; at 1536 B per buffer (one descriptor per frame) plus `CONFIG_ETH_TRANSMIT_MUTEX`, it runs
-clean. Both are bench-isolated, and ring COUNT is not the lever: 30 descriptors ran no cleaner than
+1512 B frame spanned three descriptors and a 10-descriptor ring held ~3.3 frames while the driver fires 132 back-to-back. At that depth the S31 refused ~19 000 frames and wedged twice inside 20 minutes; at 1536 B per buffer (one descriptor per frame) plus `CONFIG_ETH_TRANSMIT_MUTEX`, it runs clean. Both are bench-isolated, and ring COUNT is not the lever: 30 descriptors ran no cleaner than
 10. Cost: ~20 KB of internal DMA RAM, since the size applies to both rings
 ([lessons.md](../work/past/lessons.md)).
 
 **Static RAM: 0 B.** The driver's 1 512 B packet buffer is a class member, so it costs nothing on a
-board that never adds the driver; `check_footprint --module PanelCardDriver --firmware esp32s31`
-reports ~3 500 B of flash and no static RAM.
+board that never adds the driver; `check_footprint --module PanelCardDriver --firmware esp32s31` reports ~3 500 B of flash and no static RAM.
 
-No scenario contract yet: the driver needs a receiver card on the wire, so the numbers above are a
-bench record rather than an asserted ceiling.
+No scenario contract yet: the driver needs a receiver card on the wire, so the numbers above are a bench record rather than an asserted ceiling.
 
 ## HTTP cost of the P4's WiFi co-processor (`esp32p4rev1-eth-wifi`)
 
diff --git a/docs/reference/testing.md b/docs/reference/testing.md
index 72abbb21..9a6ca166 100644
--- a/docs/reference/testing.md
+++ b/docs/reference/testing.md
@@ -2,8 +2,8 @@
 
 What we test and how. The detailed inventory of every test lives in two auto-generated files:
 
-- **[Unit tests](tests/unit-tests.md)** — one row per `TEST_CASE`, grouped by module. Generated from `test/unit/{core,light}/unit_*.cpp`.
-- **[Scenario tests](tests/scenario-tests.md)** — one section per scenario JSON, grouped by module. Generated from `test/scenarios/{core,light}/scenario_*.json`. Each scenario runs in two tiers (in-process + live) unless flagged `live_only` (construct-mode scenarios are also skipped on live, since `main.cpp` owns the live shape).
+- **[Unit tests](tests/unit-tests.md)**: one row per `TEST_CASE`, grouped by module. Generated from `test/unit/{core,light}/unit_*.cpp`.
+- **[Scenario tests](tests/scenario-tests.md)**: one section per scenario JSON, grouped by module. Generated from `test/scenarios/{core,light}/scenario_*.json`. Each scenario runs in two tiers (in-process + live) unless flagged `live_only` (construct-mode scenarios are also skipped on live, since `main.cpp` owns the live shape).
 
 Both are produced by `moondeck/docs/generate_test_docs.py`; the source of truth is the test files themselves (see [Adding Tests](#adding-tests) below).
 
@@ -11,9 +11,9 @@ Both are produced by `moondeck/docs/generate_test_docs.py`; the source of truth
 
 Three test categories, each with a clear purpose:
 
-- **Unit tests** (desktop, `test/unit/{core,light}/unit_*.cpp`) — exercise individual MoonModules in isolation with doctest. Each test file declares `// @module <Name>` so it's categorised under that module in the generated inventory. Run via `ctest` or `./build/<host>/test/mm_tests` (`<host>` = `macos`/`linux`/`windows`, the per-host build dir). Verify a module's API, edge cases, and output independent of how it's wired into a pipeline.
-- **In-process scenarios** (desktop, `test/scenarios/{core,light}/scenario_*.json`) — exercise the system as an integrated pipeline. Each scenario is a declarative JSON file with a sequence of steps (`add_module`, `set_control`, `measure`) and optional performance bounds. The scenario runner (`test/scenario_runner.cpp`) replays the steps in-process and reports tick + heap per `measure` step. Same JSON files run against a live device through the HTTP API — that's the next tier.
-- **Live scenarios** — the same scenarios driven against a running device over REST. See [Live scenarios](#live-scenarios) below.
+- **Unit tests** (desktop, `test/unit/{core,light}/unit_*.cpp`), exercise individual MoonModules in isolation with doctest. Each test file declares `// @module <Name>` so it's categorised under that module in the generated inventory. Run via `ctest` or `./build/<host>/test/mm_tests` (`<host>` = `macos`/`linux`/`windows`, the per-host build dir). Verify a module's API, edge cases, and output independent of how it's wired into a pipeline.
+- **In-process scenarios** (desktop, `test/scenarios/{core,light}/scenario_*.json`), exercise the system as an integrated pipeline. Each scenario is a declarative JSON file with a sequence of steps (`add_module`, `set_control`, `measure`) and optional performance bounds. The scenario runner (`test/scenario_runner.cpp`) replays the steps in-process and reports tick + heap per `measure` step. Same JSON files run against a live device through the HTTP API, that's the next tier.
+- **Live scenarios**: the same scenarios driven against a running device over REST. See [Live scenarios](#live-scenarios) below.
 
 **Picking a tier for a new test.** When the behavior you want to pin only makes sense with modules wired together (e.g. "the pipeline reallocates cleanly when the grid resizes," "Drivers correctly hands the source buffer through after a child swap"), reach for a scenario first: that is what scenarios are *for*. When the behavior lives inside a single module (one function's contract, one edge case, one bug regression on a small surface), a unit test is the cheaper and faster fit. Don't extend the scenario runner with new predicates just to migrate an existing unit test, which is adding abstraction without an active need. Add predicates when a *new* scenario you're writing needs them.
 
@@ -21,16 +21,15 @@ Three test categories, each with a clear purpose:
 
 **Performance checks** verify architectural rules at runtime:
 
-- **Zero-allocation render loop** — N frames, intercept `malloc` / `free`, fail if any allocation occurs during steady-state rendering.
-- **Frame time bounds** — scenarios include `"bounds": {"fps": {"min_pct": N}}` (live, relative to baseline), `"bounds": {"fps": {"min": N}}` (absolute), or `"bounds": {"fps": {"min_fps_led_product": N}}` (throughput floor that scales to grid size).
-- **Heap-delta bounds** — scenarios include `"bounds": {"heap": {"max_delta_bytes": N}}` on a measure step to assert this step grew the heap by no more than N bytes vs the previous measure. On desktop `freeHeap()` returns 0 (unlimited), so the assertion is effectively a no-op there; it lights up on ESP32 where heap is real.
+- **Zero-allocation render loop**: N frames, intercept `malloc` / `free`, fail if any allocation occurs during steady-state rendering.
+- **Frame time bounds**: scenarios include `"bounds": {"fps": {"min_pct": N}}` (live, relative to baseline), `"bounds": {"fps": {"min": N}}` (absolute), or `"bounds": {"fps": {"min_fps_led_product": N}}` (throughput floor that scales to grid size).
+- **Heap-delta bounds**: scenarios include `"bounds": {"heap": {"max_delta_bytes": N}}` on a measure step to assert this step grew the heap by no more than N bytes vs the previous measure. On desktop `freeHeap()` returns 0 (unlimited), so the assertion is effectively a no-op there; it lights up on ESP32 where heap is real.
 
 Per-module timing, memory, and sizeof measurements per platform live in [performance.md](performance.md).
 
 ## Static analysis
 
-Tests pin behavior that runs; static analysis catches what never gets exercised. Five layers,
-cheapest first — each catches something the layer above cannot:
+Tests pin behavior that runs; static analysis catches what never gets exercised. Five layers, cheapest first, each catches something the layer above cannot:
 
 | # | Layer | Catches | Where |
 |---|---|---|---|
@@ -40,18 +39,10 @@ cheapest first — each catches something the layer above cannot:
 | 3 | RTSan + `[[clang::nonblocking]]` | allocation/blocking in the render path, **transitively** | Compile time + CI |
 | 4 | [CodeQL](../.github/codeql-config.yml) | untrusted input, whole-program taint, use-after-free | CI, Security tab |
 
-Alongside them: **lizard** counts complexity per commit for the trend (a fuzzy tokenizer, not a
-parser — it can never express an architectural rule), **clang-query** is the home for bespoke
-AST rules we invent, and the Python checks in `moondeck/check/` cover contracts whose other half
-is a Markdown page, a JSON catalog or a built binary. Every one has a MoonDeck card
-([MoonDeck.md](../moondeck/MoonDeck.md)).
+Alongside them: **lizard** counts complexity per commit for the trend (a fuzzy tokenizer, not a parser, it can never express an architectural rule), **clang-query** is the home for bespoke AST rules we invent, and the Python checks in `moondeck/check/` cover contracts whose other half is a Markdown page, a JSON catalog or a built binary. Every one has a MoonDeck card ([MoonDeck.md](../moondeck/MoonDeck.md)).
 
 **Every one of these is a report, not a gate.** They state what they find; a consumer decides
-what to do about it. `WarningsAsErrors` is empty, CodeQL never runs on `pull_request`, and the
-hot-path check never fails the event. A gate nobody can satisfy gets disabled rather than
-obeyed, and it pushes people to suppress a finding under time pressure — which is the opposite
-of why the tool is there. The exception is layer 0: compiler warnings ARE `-Werror`, because
-they are few, actionable, and fixed at the moment they appear.
+what to do about it. `WarningsAsErrors` is empty, CodeQL never runs on `pull_request`, and the hot-path check never fails the event. A gate nobody can satisfy gets disabled rather than obeyed, and it pushes people to suppress a finding under time pressure, which is the opposite of why the tool is there. The exception is layer 0: compiler warnings ARE `-Werror`, because they are few, actionable, and fixed at the moment they appear.
 
 **One rule, one owner.** Each rule is enforced in exactly one place, so a finding has one home
 and cannot drift between two tools that half-agree:
@@ -69,8 +60,7 @@ and cannot drift between two tools that half-agree:
 
 ### Verify a zero before believing it
 
-An analyser reporting "0 findings" is indistinguishable from one that read nothing, and every
-silent-failure mode below produced a plausible clean report rather than an error:
+An analyzer reporting "0 findings" is indistinguishable from one that read nothing, and every silent-failure mode below produced a plausible clean report rather than an error:
 
 | Defect | What it looked like |
 |---|---|
@@ -82,15 +72,9 @@ silent-failure mode below produced a plausible clean report rather than an error
 | Missing `-isysroot` | Under-report, not an error: 5 matches where there were 14 |
 | A baseline keyed on a name the tool no longer emits | Pins nothing while looking green (see backlog-core.md § lizard) |
 
-So: **run a control check that MUST fire.** After reaching 0, enabling a deliberately-disabled
-check (`--check readability-magic-numbers`) returned 3,307 — that is what proves the pipeline
-reads the code. `check_clang_tidy.py` also refuses to report when more than ten files fail to
-compile, because "most files errored" is a broken run, not a result.
+So: **run a control check that MUST fire.** After reaching 0, enabling a deliberately-disabled check (`--check readability-magic-numbers`) returned 3,307, that is what proves the pipeline reads the code. `check_clang_tidy.py` also refuses to report when more than ten files fail to compile, because "most files errored" is a broken run, not a result.
 
-A tool's *default* output is not its verdict, either. Evaluating an analyser without a curated
-check list measures the configuration, not the tool: an unconfigured clang-tidy run here gave
-6,073 findings, two thirds of them from one check that every comparable project disables by
-name. The curated config takes it to 30.
+A tool's *default* output is not its verdict, either. Evaluating an analyzer without a curated check list measures the configuration, not the tool: an unconfigured clang-tidy run here gave 6,073 findings, two thirds of them from one check that every comparable project disables by name. The curated config takes it to 30.
 
 ### A passing test is not evidence until it can fail
 
@@ -108,7 +92,7 @@ A green result means something only if the test could have gone red, and an agen
 
 Two principles drive every standard below:
 
-1. **The test file is the source of truth.** Every fact about a test (which module, which scenario, what each case verifies, what each step does) lives in the test file itself. The generated docs and MoonDeck views read from there — they don't get edited by hand.
+1. **The test file is the source of truth.** Every fact about a test (which module, which scenario, what each case verifies, what each step does) lives in the test file itself. The generated docs and MoonDeck views read from there, they don't get edited by hand.
 2. **One pattern, easy to spot in 30 seconds.** Every unit test follows the same header shape. Every scenario follows the same JSON shape. A new contributor recognises an existing test, then writes the next one by analogy.
 
 ### File layout
@@ -127,44 +111,44 @@ test/
 └── js/            <topic>.test.mjs                    # host-side: mooninstaller logic
 ```
 
-A test lives under the subfolder of its **primary** `@module`'s source domain (e.g. `Layer` lives in `src/light/`, so `unit_Layer_extrude.cpp` goes in `test/unit/light/`). Cross-domain awareness travels through the `@also` list, not the directory. There's no `platform/` subfolder today — `src/platform/` is a pure abstraction layer whose desktop backend every unit test implicitly exercises; ESP32 platform code never runs on the desktop, so there's nothing to put there yet.
+A test lives under the subfolder of its **primary** `@module`'s source domain (e.g. `Layer` lives in `src/light/`, so `unit_Layer_extrude.cpp` goes in `test/unit/light/`). Cross-domain awareness travels through the `@also` list, not the directory. There's no `platform/` subfolder today, `src/platform/` is a pure abstraction layer whose desktop backend every unit test implicitly exercises; ESP32 platform code never runs on the desktop, so there's nothing to put there yet.
 
 ### Host-side tests (Python + JS)
 
-The C++ `ctest` / scenario suites can't reach the **Python** (MoonDeck, build scripts) or **JS** (web installer) code, so those get their own host-side unit tier — `test/python/` (pytest, run `uv run --with pytest --with pyserial pytest test/python`) and `test/js/` (Node's built-in runner, run `node --test "test/js/**/*.test.mjs"`; no `package.json`/`npm install`). Both run in `.github/workflows/test.yml` on every PR and are commit gates (CLAUDE.md Event 1, gate 10) when `moondeck/` / `mooninstaller/` / the test dirs change. Python test files carry their deps in a PEP-723 `# /// script` block (the repo's convention — there's no central `pyproject.toml`); pyserial is a dep only because `improv_provision.py`'s import guard exits without it, not because the frame logic needs it.
+The C++ `ctest` / scenario suites can't reach the **Python** (MoonDeck, build scripts) or **JS** (web installer) code, so those get their own host-side unit tier, `test/python/` (pytest, run `uv run --with pytest --with pyserial pytest test/python`) and `test/js/` (Node's built-in runner, run `node --test "test/js/**/*.test.mjs"`; no `package.json`/`npm install`). Both run in `.github/workflows/test.yml` on every PR and are commit gates (CLAUDE.md Event 1, gate 10) when `moondeck/` / `mooninstaller/` / the test dirs change. Python test files carry their deps in a PEP-723 `# /// script` block (the repo's convention, there's no central `pyproject.toml`); pyserial is a dep only because `improv_provision.py`'s import guard exits without it, not because the frame logic needs it.
 
 #### What's covered today: the Improv frame wire format
 
-The Improv serial frame is implemented **three times** — device C++ (`src/core/ImprovFrame.h`), Python (`moondeck/build/improv_provision.py`), and installer JS (`mooninstaller/improv-frame.js`) — so a drift in any one silently breaks provisioning. Both suites assert the **same golden vector** so the Python and JS builders provably agree (hand-verified against the C++ sum-mod-256 checksum):
+The Improv serial frame is implemented **three times**, device C++ (`src/core/ImprovFrame.h`), Python (`moondeck/build/improv_provision.py`), and installer JS (`mooninstaller/improv-frame.js`), so a drift in any one silently breaks provisioning. Both suites assert the **same golden vector** so the Python and JS builders provably agree (hand-verified against the C++ sum-mod-256 checksum):
 
 ```
 buildImprovFrame(type=0x03, payload=[0x01])  ==  49 4d 50 52 4f 56 01 03 01 01 e3
                                                   └IMPROV┘ v  t  l  p  checksum
 ```
 
-**`test/python/test_improv_frame.py`** (pytest) — pins the *shared envelope* that MoonDeck and the provisioning scripts speak (`build_frame` / `checksum` in `improv_provision.py`):
+**`test/python/test_improv_frame.py`** (pytest), pins the *shared envelope* that MoonDeck and the provisioning scripts speak (`build_frame` / `checksum` in `improv_provision.py`):
 
 | Test | Pins |
 |---|---|
-| `test_checksum_is_sum_mod_256` | checksum is `sum(bytes) & 0xFF` — empty → 0, `0xFF 0xFF` → `0xFE` (wraps) |
+| `test_checksum_is_sum_mod_256` | checksum is `sum(bytes) & 0xFF`, empty → 0, `0xFF 0xFF` → `0xFE` (wraps) |
 | `test_frame_layout` | `IMPROV` magic, version 1, type, length, payload, total length |
 | `test_golden_vector_g1` | the exact 11-byte golden frame above (the cross-impl anchor) |
 | `test_checksum_covers_header_through_payload` | checksum spans header→payload, excludes the checksum byte itself |
 
-**`test/js/improv-frame.test.mjs`** (node:test) — pins the same envelope *plus* the `APPLY_OP` chunking that only the installer JS and device C++ implement (Python's provisioning path does WIFI_SETTINGS, not config push):
+**`test/js/improv-frame.test.mjs`** (node:test), pins the same envelope *plus* the `APPLY_OP` chunking that only the installer JS and device C++ implement (Python's provisioning path does WIFI_SETTINGS, not config push):
 
 | Test | Pins |
 |---|---|
 | frame layout | magic / version / type / length / payload / checksum positions |
 | checksum is sum-mod-256 | matches Python + C++ |
 | golden vector G1 | the same 11-byte frame as the pytest anchor |
-| G2 — small APPLY_OP `set` is a single frame | `[0xFC][seq=0][last=1]` header + the op JSON byte-identical in the payload |
-| G3 — a >125-byte op chunks into ordered frames | two frames, `seq` 0→1, `last` 0→1, the 125-byte chunk boundary, reassembly reproduces the op JSON exactly |
+| G2, small APPLY_OP `set` is a single frame | `[0xFC][seq=0][last=1]` header + the op JSON byte-identical in the payload |
+| G3, a >125-byte op chunks into ordered frames | two frames, `seq` 0→1, `last` 0→1, the 125-byte chunk boundary, reassembly reproduces the op JSON exactly |
 | at-least-one-frame | an empty op still emits one frame with `last=1` (so `last` always sends) |
 
-The JS suite proves the installer *chunks* an op correctly; the **device side that reassembles those chunks** is pinned by the C++ `unit_ImprovOpReassembler` suite (`src/core/ImprovOpReassembler.h`, the pure state machine behind the device's `APPLY_OP` handler — extracted from `platform_esp32_improv.cpp` so it's desktop-testable). It covers the full receive contract: in-order multi-chunk reassembly + NUL-termination, **duplicate-chunk rejection** and **out-of-order/skipped-seq rejection** (the guard against an installer retry corrupting the buffer), **overflow** rejection at the buffer-minus-NUL boundary, mid-stream `seq 0` abandoning a partial op, and clean recovery after every error. Encode (JS) + reassemble (C++) together prove APPLY_OP end to end without hardware.
+The JS suite proves the installer *chunks* an op correctly; the **device side that reassembles those chunks** is pinned by the C++ `unit_ImprovOpReassembler` suite (`src/core/ImprovOpReassembler.h`, the pure state machine behind the device's `APPLY_OP` handler, extracted from `platform_esp32_improv.cpp` so it's desktop-testable). It covers the full receive contract: in-order multi-chunk reassembly + NUL-termination, **duplicate-chunk rejection** and **out-of-order/skipped-seq rejection** (the guard against an installer retry corrupting the buffer), **overflow** rejection at the buffer-minus-NUL boundary, mid-stream `seq 0` abandoning a partial op, and clean recovery after every error. Encode (JS) + reassemble (C++) together prove APPLY_OP end to end without hardware.
 
-**`test/python/test_installer_manifests.py`** (pytest) — pins the web installer's per-release file contract. For every `ships: true` firmware in `mooninstaller/firmwares.json` it runs `moondeck/build/generate_manifest.py` (with a synthetic `flasher_args.json`, so no firmware build is needed) and asserts the manifest is valid (a `chipFamily` + non-empty `parts[]`) AND that **every part filename matches one of the globs the release workflow stages onto Pages** (`firmware-*.bin` / `shared-*.bin` / `partition-table-*.bin` — `shared-*.bin` covers the plain ota-data plus a MoonBase manifest's `shared-ota-data-slot0.bin` and `shared-moonbase-<chip>.bin`). A manifest that names a file outside those globs points at something the deploy never stages → the installer 404s at fetch-firmware (the failure that shipped a broken v2.0.0 installer). The test guards the manifest-generation ↔ staged-files contract; the *deploy mechanics* that stage them (per-tag, in `release.yml`) are workflow shell logic a unit test can't reach, so the two are complementary.
+**`test/python/test_installer_manifests.py`** (pytest), pins the web installer's per-release file contract. For every `ships: true` firmware in `mooninstaller/firmwares.json` it runs `moondeck/build/generate_manifest.py` (with a synthetic `flasher_args.json`, so no firmware build is needed) and asserts the manifest is valid (a `chipFamily` + non-empty `parts[]`) AND that **every part filename matches one of the globs the release workflow stages onto Pages** (`firmware-*.bin` / `shared-*.bin` / `partition-table-*.bin`, `shared-*.bin` covers the plain ota-data plus a MoonBase manifest's `shared-ota-data-slot0.bin` and `shared-moonbase-<chip>.bin`). A manifest that names a file outside those globs points at something the deploy never stages → the installer 404s at fetch-firmware (the failure that shipped a broken v2.0.0 installer). The test guards the manifest-generation ↔ staged-files contract; the *deploy mechanics* that stage them (per-tag, in `release.yml`) are workflow shell logic a unit test can't reach, so the two are complementary.
 
 **`test/python/test_check_specs_drift.py`** (pytest) pins the two spec-drift guards in `moondeck/check/check_specs.py` (the spec-check commit gate). Some facts live in both the `.h` and the module doc in different forms: a control's **numeric range** (`addControl("floor", floor, 0, 255)` vs the prose "noise floor (0–255)") and the **author/source URL** (`// Author: …, <url>` vs the `Origin:` markdown link). Neither can be single-sourced, they're the same fact for two audiences, so instead the gate *validates* them: if the doc restates a control's range and it conflicts with the `.h`, or an `.h` author URL is missing from the doc, the spec check flags it. The checks are block-scoped on the consolidated catalog pages (a control name shared across modules, `fps`/`fadeRate`, matches only its own module's block), and tolerant of the human range spellings (`1–8` / `1-8` / `1 to 8`): this suite pins both the catch and the no-false-alarm behavior.
 
@@ -172,8 +156,8 @@ MoonDeck's pure logic (catalog reverse-lookup, state migration) and the installe
 
 ### Naming convention
 
-- **Unit tests:** `unit_<ExactModuleName>[_<topic>].cpp` — `<ExactModuleName>` is the **CamelCase** class name as it appears in `// @module` (and in the source: `Layer`, `MoonModule`, `MultiplyModifier`, `NetworkSendDriver`). The optional `<topic>` collapses when the file's the only test for its module (`unit_Color.cpp` is fine if `@module Color`); add it when one module has several test files (`unit_Layer_extrude.cpp`, `unit_Layer_zero_grid.cpp`, …) or when the topic genuinely clarifies what the file covers (`unit_FilesystemModule_persistence.cpp`).
-- **Scenarios:** `scenario_<ExactModuleName>_<topic>.json` — same module-naming rule; the topic is always present because scenarios always cross multiple modules and the topic distinguishes the focus.
+- **Unit tests:** `unit_<ExactModuleName>[_<topic>].cpp`, `<ExactModuleName>` is the **CamelCase** class name as it appears in `// @module` (and in the source: `Layer`, `MoonModule`, `MultiplyModifier`, `NetworkSendDriver`). The optional `<topic>` collapses when the file's the only test for its module (`unit_Color.cpp` is fine if `@module Color`); add it when one module has several test files (`unit_Layer_extrude.cpp`, `unit_Layer_zero_grid.cpp`, …) or when the topic genuinely clarifies what the file covers (`unit_FilesystemModule_persistence.cpp`).
+- **Scenarios:** `scenario_<ExactModuleName>_<topic>.json`, same module-naming rule; the topic is always present because scenarios always cross multiple modules and the topic distinguishes the focus.
 - The **`"name"` field inside each scenario JSON** matches the filename stem exactly (e.g. `"name": "scenario_Layer_base_pipeline"`). The runner, the MoonDeck dropdown, the generated docs and `--name` on the CLI all use this single identifier.
 
 ### Unit-test file shape
@@ -205,7 +189,7 @@ Header annotations recognised by the parser:
 |-----|-----------|-------|
 | `// @module <Name>` | **yes** | Exactly one. CamelCase, must match a real module class. |
 | `// @also <A>, <B>, ...` | optional | Comma-separated peer modules this test also exercises. Surfaces as “Also touches” in the generated docs and lets the MoonDeck module filter include the test for either domain. |
-| `// @description <one-line>` | optional | Short file-level summary. Multiple `@description` lines concatenate. Use only when the tests don't speak for themselves — usually omit. |
+| `// @description <one-line>` | optional | Short file-level summary. Multiple `@description` lines concatenate. Use only when the tests don't speak for themselves, usually omit. |
 
 Per-`TEST_CASE` description rules:
 
@@ -216,8 +200,8 @@ Per-`TEST_CASE` description rules:
 
 Every scenario carries a top-level `mode` field that says what shape the scenario expects the world to be in. Two values:
 
-- **`"mode": "construct"`** — the scenario builds the pipeline from an empty scheduler. Lots of `add_module` steps; the first `measure` happens after everything is wired. **Runs in-process only.** The live device's top-level shape is policy-fixed in `main.cpp` (see [src/core/HttpServerModule.cpp:639](../src/core/HttpServerModule.cpp#L639) — `/api/modules` rejects top-level adds), so "build from scratch" can't happen on a live device without re-flashing. The live runner skips construct scenarios with a clear note.
-- **`"mode": "mutate"`** — the scenario assumes a wired pipeline and tweaks it (`set_control` heavy). Runs in both tiers. The in-process runner replays an embedded **`fixture`** array (same shape as `steps`, but all `add_module`) that builds the same pipeline `main.cpp` does, then runs the actual steps. The live runner skips the fixture (device is its own fixture) and pre-flights that every id the steps touch is actually present on the device — a missing id is a hard fail, not a silent skip.
+- **`"mode": "construct"`**: the scenario builds the pipeline from an empty scheduler. Lots of `add_module` steps; the first `measure` happens after everything is wired. **Runs in-process only.** The live device's top-level shape is policy-fixed in `main.cpp` (see [src/core/HttpServerModule.cpp:639](../src/core/HttpServerModule.cpp#L639), `/api/modules` rejects top-level adds), so "build from scratch" can't happen on a live device without re-flashing. The live runner skips construct scenarios with a clear note.
+- **`"mode": "mutate"`**: the scenario assumes a wired pipeline and tweaks it (`set_control` heavy). Runs in both tiers. The in-process runner replays an embedded **`fixture`** array (same shape as `steps`, but all `add_module`) that builds the same pipeline `main.cpp` does, then runs the actual steps. The live runner skips the fixture (device is its own fixture) and pre-flights that every id the steps touch is actually present on the device, a missing id is a hard fail, not a silent skip.
 
 Picking the right mode:
 - If your scenario starts with empty Layouts/Layer/Drivers wiring, it's **construct**. It will not run live.
@@ -229,7 +213,7 @@ A `mutate` scenario that needs platform-bound modules (Network mDNS, WiFi, OTA)
 
 ### Reset block: idempotent scenarios
 
-`mutate` scenarios mutate shared controls (Grid size, Multiply mirror toggles, Preview detail). Without restoring those controls to known values at scenario start, measurements become coupled to *which other scenarios ran first* — last scenario's leftover state poisons this one's baseline. The fix is a top-level **`reset`** array, an `add_module`/`set_control`-shape list that runs *before* the scenario's own `steps`:
+`mutate` scenarios mutate shared controls (Grid size, Multiply mirror toggles, Preview detail). Without restoring those controls to known values at scenario start, measurements become coupled to *which other scenarios ran first*, last scenario's leftover state poisons this one's baseline. The fix is a top-level **`reset`** array, an `add_module`/`set_control`-shape list that runs *before* the scenario's own `steps`:
 
 ```json
 "reset": [
@@ -238,13 +222,13 @@ A `mutate` scenario that needs platform-bound modules (Network mDNS, WiFi, OTA)
 ],
 ```
 
-Both runners replay reset before baseline. The baseline measurement is taken *after* reset, so it reflects the normalised pipeline state. Reset steps don't have `measure: true` — they're prep, not assertions.
+Both runners replay reset before baseline. The baseline measurement is taken *after* reset, so it reflects the normalized pipeline state. Reset steps don't have `measure: true`, they're prep, not assertions.
 
 Convention: reset every control your scenario writes, plus any production-default control its `steps` implicitly depend on. The cost is a few extra steps; the payoff is scenarios you can run in any order and `contract` values you can trust.
 
 ### Performance contracts (`contract[<target>]`)
 
-Every measurable step carries a per-target `contract` block — the **performance contract** projectMM commits to delivering on that platform. The runner compares each measurement to the contract and fails if the device misses it.
+Every measurable step carries a per-target `contract` block, the **performance contract** projectMM commits to delivering on that platform. The runner compares each measurement to the contract and fails if the device misses it.
 
 ```json
 "contract": {
@@ -265,19 +249,19 @@ Every measurable step carries a per-target `contract` block — the **performanc
 
 **Contract semantics:**
 
-- `tick_us` is a **ceiling** — the device must deliver this tick or faster. Going faster than contract is good news, never a failure.
-- `free_heap` is a **floor** — the device must deliver at least this much free heap. Growing heap is good news.
-- Both are **hand-set promises**, not auto-captured last readings. Renegotiating a contract requires `--update-contract --reason "..."` — see below.
+- `tick_us` is a **ceiling**, the device must deliver this tick or faster. Going faster than contract is good news, never a failure.
+- `free_heap` is a **floor**, the device must deliver at least this much free heap. Growing heap is good news.
+- Both are **hand-set promises**, not auto-captured last readings. Renegotiating a contract requires `--update-contract --reason "..."`, see below.
 - `set_by` records when the contract was last (re)negotiated; `reason` records why. Both stamped automatically by `--update-contract`.
 
 Target keys match `SystemModule.firmware` on a flashed device (`esp32`, `esp32-eth`, `esp32-eth-wifi`, `esp32s3-n16r8`, …) plus `desktop-macos` / `desktop-linux` / `desktop-windows` for desktop builds. The in-process runner picks the host OS automatically; the live runner reads the device's `firmware` control. (See [MoonInstaller, firmware vs deviceModel vs board](../explanation/architecture/mooninstaller.md#the-three-words) for the distinction.)
 
-**Tolerance** absorbs run-to-run jitter only — not "I don't care":
+**Tolerance** absorbs run-to-run jitter only, not "I don't care":
 
 - Default tick tolerance: **20% on desktop-*** (multi-process OS scheduling), **10% on ESP32** (lwIP / EMAC jitter).
 - Default heap tolerance: same percentages.
-- Default absolute floor on tick: **200 µs on desktop-***, **5 µs on ESP32** — prevents percentage tolerance from going sub-noise on tiny ticks.
-- **On desktop-*, the floor dominates** for any contract under ~1 ms tick (effectively all desktop contracts today: 5–300 µs). The percentage knob only matters for slow desktop tick. This is honest: a 100 µs desktop contract realistically can't assert anything tighter than ±200 µs because multi-process scheduling jitter routinely adds 100+ µs all by itself. If you need a tight assertion, run the contract on ESP32 — bounded RTOS clocks are tighter.
+- Default absolute floor on tick: **200 µs on desktop-***, **5 µs on ESP32**, prevents percentage tolerance from going sub-noise on tiny ticks.
+- **On desktop-*, the floor dominates** for any contract under ~1 ms tick (effectively all desktop contracts today: 5–300 µs). The percentage knob only matters for slow desktop tick. This is honest: a 100 µs desktop contract realistically can't assert anything tighter than ±200 µs because multi-process scheduling jitter routinely adds 100+ µs all by itself. If you need a tight assertion, run the contract on ESP32, bounded RTOS clocks are tighter.
 - Per-step overrides: `tick_tolerance_pct`, `heap_tolerance_pct`, `tolerance_us` inside the per-target block.
 
 **Predicted-vs-actual heap.** The runner additionally prints the sum of `dynamicBytes()` from the live module tree as `model=N`. This is what the system *says* it allocated; comparing it to `free_heap` reveals framework overhead (lwIP, WiFi, HTTP buffers). The model number is informational, not asserted.
@@ -305,7 +289,7 @@ Alongside `contract`, every measurable step keeps a **rolling [min, max] range**
 
 Read directions match the contract semantics: tick contract is a *ceiling*, so the **max** of the observed range is the value that could fail it; heap and block contracts are *floors*, so the **min** of the observed range is the failure-side. The `at` field carries `[first_seen, last_updated]` so a stale range stands out (e.g. "last_updated is six months old → re-sweep this target").
 
-Updated by **every** successful run of `run_scenario.py` or `run_live_scenario.py` for the active target — no flag needed. A range that never widens produces no diff. When a contract is renegotiated via `--update-contract` (see below), the observed range *resets* to the new single-point measurement because the previous range belonged to the old contract.
+Updated by **every** successful run of `run_scenario.py` or `run_live_scenario.py` for the active target, no flag needed. A range that never widens produces no diff. When a contract is renegotiated via `--update-contract` (see below), the observed range *resets* to the new single-point measurement because the previous range belonged to the old contract.
 
 Open any scenario JSON and the range tells you both the typical case (the value sits in the range) and the variance (range width). A wide gap between the max of the observed range and the contract ceiling means there's headroom to tighten the promise; a max creeping up against the ceiling means a regression may be imminent.
 
@@ -323,7 +307,7 @@ uv run moondeck/scenario/run_live_scenario.py --host 192.168.1.210 \
     --update-contract --reason "added DMX driver overhead"
 ```
 
-Both runners require `--reason`. They stamp `set_by` to today's date and write the new contract for the active target only — other targets are untouched. Review the diff carefully; tightening or loosening a contract is a deliberate act that ships to the project's public KPIs.
+Both runners require `--reason`. They stamp `set_by` to today's date and write the new contract for the active target only, other targets are untouched. Review the diff carefully; tightening or loosening a contract is a deliberate act that ships to the project's public KPIs.
 
 ### Scenario file shape
 
@@ -373,19 +357,19 @@ Every `scenario_*.json` carries top-level metadata plus a `description` per step
 }
 ```
 
-`name` matches the filename stem exactly; `module` drives the doc grouping and the MoonDeck filter; `also` lists peer modules. `mode` is `construct` or `mutate` (see above). `fixture` is required for `mutate` scenarios that need to run in-process. `reset` (optional but recommended for mutate scenarios) normalises controls before measurement so contract assertions aren't coupled to run order. `contract[<target>]` carries the per-target performance contract — see § Performance contracts. `observed[<target>]` carries a rolling `[min, max]` range per metric (tick_us, free_heap, max_alloc_block) plus `at: [first_seen_iso, last_seen_iso]` timestamps — widen-only across runs, so each run extends the range rather than replacing it; see § Persistent observations. Per-step `description` is recommended (a missing one falls back to the step's `name`/`op` in the generated view). Optional top-level `"live_only": true` keeps the scenario out of the in-process runner. Optional per-step `"measure": true` triggers a measurement after the step with optional `bounds`:
+`name` matches the filename stem exactly; `module` drives the doc grouping and the MoonDeck filter; `also` lists peer modules. `mode` is `construct` or `mutate` (see above). `fixture` is required for `mutate` scenarios that need to run in-process. `reset` (optional but recommended for mutate scenarios) normalizes controls before measurement so contract assertions aren't coupled to run order. `contract[<target>]` carries the per-target performance contract, see § Performance contracts. `observed[<target>]` carries a rolling `[min, max]` range per metric (tick_us, free_heap, max_alloc_block) plus `at: [first_seen_iso, last_seen_iso]` timestamps, widen-only across runs, so each run extends the range rather than replacing it; see § Persistent observations. Per-step `description` is recommended (a missing one falls back to the step's `name`/`op` in the generated view). Optional top-level `"live_only": true` keeps the scenario out of the in-process runner. Optional per-step `"measure": true` triggers a measurement after the step with optional `bounds`:
 
-- `bounds.fps.min` — absolute FPS floor (in-process).
-- `bounds.fps.min_pct` — relative to baseline (live runner).
-- `bounds.fps.min_fps_led_product` — throughput floor scaling to grid size.
-- `bounds.heap.max_delta_bytes` — max heap growth vs the previous measure step (lights up on ESP32; no-op on desktop).
+- `bounds.fps.min`: absolute FPS floor (in-process).
+- `bounds.fps.min_pct`: relative to baseline (live runner).
+- `bounds.fps.min_fps_led_product`: throughput floor scaling to grid size.
+- `bounds.heap.max_delta_bytes`: max heap growth vs the previous measure step (lights up on ESP32; no-op on desktop).
 
-**`bounds` and `contract` co-exist by design** — they catch different things on the same measure step:
+**`bounds` and `contract` co-exist by design**, they catch different things on the same measure step:
 
 - `bounds` are **relative within a single run**: "this step's FPS must be at least 80% of *this run's* baseline." Catches a measurement collapsing mid-scenario (effect silently dropping frames, mirror toggle leaving a degraded state). Cheap because it doesn't depend on history.
 - `contract.<target>` is an **absolute promise across runs**: "this step must hit ≤ X µs / ≥ Y bytes heap on this platform." Catches regressions over time that bounds can't see (the baseline itself drifts).
 
-If you only need one, prefer `contract` — it's the load-bearing assertion. `bounds` is the within-run sanity check.
+If you only need one, prefer `contract`, it's the load-bearing assertion. `bounds` is the within-run sanity check.
 
 Unknown JSON keys are ignored by both runners (C++ and Python), so adding a new field is safe.
 
@@ -407,7 +391,7 @@ uv run moondeck/docs/generate_test_docs.py            # writes both docs
 uv run moondeck/docs/generate_test_docs.py --check    # exits non-zero on drift (CI-friendly)
 ```
 
-`--check` exits non-zero when the docs are stale — a contributor who adds a `TEST_CASE` without re-running the generator gets flagged. Run it in CI or before a commit to keep the generated docs in sync.
+`--check` exits non-zero when the docs are stale, a contributor who adds a `TEST_CASE` without re-running the generator gets flagged. Run it in CI or before a commit to keep the generated docs in sync.
 
 ## Unit tests
 
@@ -442,16 +426,16 @@ uv run moondeck/scenario/run_scenario.py --module Layer                       #
 
 Or via MoonDeck (Desktop tab → Scenarios card). The module dropdown is shared with the Unit Test card above it: pick a module once and both card's run-set narrows. Steps button shows the per-scenario step list.
 
-**Scenarios with `"live_only": true`** are skipped in-process (they need a real device — e.g. measure real-Ethernet throughput at a given grid size). They appear in the inventory but only run via the live tier below. The symmetric case (scenarios that only make sense in-process) is already handled by `mode: "construct"`, which skips on live.
+**Scenarios with `"live_only": true`** are skipped in-process (they need a real device, e.g. measure real-Ethernet throughput at a given grid size). They appear in the inventory but only run via the live tier below. The symmetric case (scenarios that only make sense in-process) is already handled by `mode: "construct"`, which skips on live.
 
 **Step ops** the in-process runner understands today:
 
-- `add_module` — instantiate a module by `type`, register it under `id`. Top-level when no `parent_id`. Mid-scenario adds (after the first `measure` step) run setup() + prepareTree() immediately, mirroring the live `/api/modules` POST shape.
-- `remove_module` / `delete_module` — delete a child module from its parent (release + recursive free + prepareTree). The two names are aliases (the in-process and live runners must never diverge on op names, or a scenario silently no-ops on one tier). Refuses non-editable submodules (`userEditable()==false`) and top-level modules.
-- `replace_module` — swap a child for a fresh module of another `type` at the same slot. A default-named module relabels to the new type; a custom/scenario id is preserved so later steps can still address it. Mirrors `/api/modules/<name>/replace`.
-- `clear_children` — delete every deletable child of a container (`id`), leaving the container. The "prepare my own canvas" primitive: a scenario assumes nothing about the device's starting tree, clears a container, then adds what it needs. Non-editable children (Board, Preview, Improv) are skipped.
-- `set_control` — write a control on an already-added module (`id` + `key` + `value`). Mirrors `handleSetControl`: applies the typed write, calls `onControlChanged()`, and triggers `Scheduler::prepareTree()` if `affectsPrepare` returns true. Today supports Uint8 / Uint16 / Int16 / Bool / Text / Password / Select. A step may carry `"optional": true` — a best-effort write (e.g. shrink a grid that may not exist) that's skipped, not failed, when the target is absent.
-- `measure` — pure measurement step. Runs warmup + measure frames, prints per-step tick / FPS / lights / heap-delta, applies any `bounds` assertions for this step.
+- `add_module`: instantiate a module by `type`, register it under `id`. Top-level when no `parent_id`. Mid-scenario adds (after the first `measure` step) run setup() + prepareTree() immediately, mirroring the live `/api/modules` POST shape.
+- `remove_module` / `delete_module`, delete a child module from its parent (release + recursive free + prepareTree). The two names are aliases (the in-process and live runners must never diverge on op names, or a scenario silently no-ops on one tier). Refuses non-editable submodules (`userEditable()==false`) and top-level modules.
+- `replace_module`: swap a child for a fresh module of another `type` at the same slot. A default-named module relabels to the new type; a custom/scenario id is preserved so later steps can still address it. Mirrors `/api/modules/<name>/replace`.
+- `clear_children`: delete every deletable child of a container (`id`), leaving the container. The "prepare my own canvas" primitive: a scenario assumes nothing about the device's starting tree, clears a container, then adds what it needs. Non-editable children (Board, Preview, Improv) are skipped.
+- `set_control`: write a control on an already-added module (`id` + `key` + `value`). Mirrors `handleSetControl`: applies the typed write, calls `onControlChanged()`, and triggers `Scheduler::prepareTree()` if `affectsPrepare` returns true. Today supports Uint8 / Uint16 / Int16 / Bool / Text / Password / Select. A step may carry `"optional": true`, a best-effort write (e.g. shrink a grid that may not exist) that's skipped, not failed, when the target is absent.
+- `measure`: pure measurement step. Runs warmup + measure frames, prints per-step tick / FPS / lights / heap-delta, applies any `bounds` assertions for this step.
 
 A step can also set `"measure": true` on a non-measure op (e.g. mark the last `add_module` as the one to measure after); the runner treats either shape identically.
 
@@ -470,13 +454,13 @@ MoonDeck's Live tab wraps the same workflow: the Network bar at the top selects
 
 ![MoonDeck Live tab](../assets/ui/moondeck_live.png)
 
-All scenarios use relative FPS bounds (`min_pct`) so they pass on any device — desktop at 10K FPS or ESP32 at 17 FPS. Settle time is 3 seconds to let the pipeline stabilise after rebuilds.
+All scenarios use relative FPS bounds (`min_pct`) so they pass on any device, desktop at 10K FPS or ESP32 at 17 FPS. Settle time is 3 seconds to let the pipeline stabilise after rebuilds.
 
 Scenarios that add modules (e.g. `scenario_Layer_base_pipeline`, `scenario_Layer_memory_1to1`) create temporary modules on the running device and clean them up at the end (`- Rainbow (cleanup)`). Modules that already exist show `=` instead of `+`.
 
 Memory tracking works on ESP32: `freeHeap` and `freeInternalHeap` report real values. Desktop returns 0 (unlimited). The control-change scenario verifies no memory leaks by checking that heap returns to baseline after a mirror toggle.
 
-One live-tier test lives outside the scenario JSON schema because it spans **multiple devices**: `uv run moondeck/scenario/run_network_live.py` runs a lights-over-UDP matrix (ArtNet, E1.31 and DDP) over every online board in moondeck.json — each board is once the sender, all others listen, and reception is asserted by reading each device's `/ws` preview stream (see [MoonDeck.md § run_network_live](../moondeck/MoonDeck.md#run_network_live)). A device matrix needs loops and per-round state the declarative scenario JSON can't express, so it follows the `improv_smoke_test.py` script shape instead.
+One live-tier test lives outside the scenario JSON schema because it spans **multiple devices**: `uv run moondeck/scenario/run_network_live.py` runs a lights-over-UDP matrix (ArtNet, E1.31 and DDP) over every online board in moondeck.json, each board is once the sender, all others listen, and reception is asserted by reading each device's `/ws` preview stream (see [MoonDeck.md § run_network_live](../moondeck/MoonDeck.md#run_network_live)). A device matrix needs loops and per-round state the declarative scenario JSON can't express, so it follows the `improv_smoke_test.py` script shape instead.
 
 ## UI scenarios
 
@@ -497,19 +481,19 @@ The runs live in `test/uiscenarios/clips/`, the engine in `moondeck/uiscenario/`
 
 All live scenarios pass on both desktop and ESP32 with `min_pct: 80` relative bounds. Per-module timing, memory allocation, and sizeof measurements for each platform are in [performance.md](performance.md).
 
-### ESP32 — classic, no PSRAM (measured on an Olimex ESP32-Gateway Rev G)
+### ESP32: classic, no PSRAM (measured on an Olimex ESP32-Gateway Rev G)
 
-- 128×128 grid (16,384 lights) — all live scenarios pass.
+- 128×128 grid (16,384 lights), all live scenarios pass.
 - Memory tracking verified: mirror toggle shows heap changes, returns to baseline (no leaks).
 - Ethernet (LAN8720 RMII) connects via DHCP.
 - Device discovery from MoonDeck finds the ESP32 on port 80.
 
 ## Adding Tests
 
-**Unit test:** add a `TEST_CASE` to the appropriate `test/unit/{core,light}/unit_<ExactModuleName>[_<topic>].cpp` file. Each file carries `// @module <ExactCamelCaseName>` at the top, plus a single `//` description line above each `TEST_CASE`. Add a new file when no existing test covers your module — pick the subfolder matching the module's `src/` domain. After adding cases, run `uv run moondeck/docs/generate_test_docs.py` so the generated inventory matches.
+**Unit test:** add a `TEST_CASE` to the appropriate `test/unit/{core,light}/unit_<ExactModuleName>[_<topic>].cpp` file. Each file carries `// @module <ExactCamelCaseName>` at the top, plus a single `//` description line above each `TEST_CASE`. Add a new file when no existing test covers your module, pick the subfolder matching the module's `src/` domain. After adding cases, run `uv run moondeck/docs/generate_test_docs.py` so the generated inventory matches.
 
 **Scenario test:** create a JSON file under `test/scenarios/{core,light}/` named `scenario_<ExactModuleName>_<topic>.json`. The top-level needs `name` (matching the filename stem), `module`, optional `also`, `description`, `mode` (`construct` or `mutate`), and optional `"live_only": true` if the scenario can only run against a real device. Each `steps[]` entry has an `op` (`add_module`, `set_control`, `measure`), a `name`, a `description` field that the doc generator picks up, optional `"measure": true` to run a measurement after the op, and optional `bounds` (`fps` and/or `heap`). The scenario runner auto-discovers all `.json` files under `test/scenarios/` recursively.
 
 **Regression test:** when fixing a bug, add a test that reproduces it. The test's description (the `//` line for unit tests, the `description` JSON field for scenarios) should mention the root cause so the connection stays traceable in the generated inventory.
 
-**Doc check:** `moondeck/docs/generate_test_docs.py --check` exits non-zero if regeneration would change `docs/tests/*.md` — a CI-friendly way to catch metadata that drifted from the source.
+**Doc check:** `moondeck/docs/generate_test_docs.py --check` exits non-zero if regeneration would change `docs/tests/*.md`, a CI-friendly way to catch metadata that drifted from the source.
diff --git a/docs/tutorials/generative-effects.md b/docs/tutorials/generative-effects.md
index 277f7c1b..b54c02eb 100644
--- a/docs/tutorials/generative-effects.md
+++ b/docs/tutorials/generative-effects.md
@@ -1,15 +1,9 @@
 # Making beautiful effects
 
-Most LED effects are written a light at a time: loop over the pixels, work out a
-color, write it. That works, and it takes a long time to get from "it lights up"
-to "I want to keep watching it".
+Most LED effects are written a light at a time: loop over the pixels, work out a color, write it. That works, and it takes a long time to get from "it lights up" to "I want to keep watching it".
 
 This page is about the other way. projectMM ships a set of **power functions**:
-the handful of algorithms that computer graphics has used for forty years to make
-things look alive. You do not implement them, you compose them. And because most of
-that vocabulary is reachable from **MoonLive** scripts as well as from compiled
-C++, a few dozen lines of script gets you an effect that would otherwise be a
-project.
+the handful of algorithms that computer graphics has used for forty years to make things look alive. You do not implement them, you compose them. And because most of that vocabulary is reachable from **MoonLive** scripts as well as from compiled C++, a few dozen lines of script gets you an effect that would otherwise be a project.
 
 > Never opened the interface? Start with **[Install & first light](../gettingstarted.md)**
 > and **[How projectMM works](how-projectmm-works.md)**, then come back.
@@ -20,29 +14,20 @@ project.
 
 Here is the shift that makes everything else easy.
 
-Instead of asking *"what should pixel 37 be?"*, ask *"what is the value of this
-field at (x, y, z), right now?"* The field is a function of position and time. You
-sample it once per light, turn the number into a color, and the picture falls out.
+Instead of asking *"what should pixel 37 be?"*, ask *"what is the value of this field at (x, y, z), right now?"* The field is a function of position and time. You sample it once per light, turn the number into a color, and the picture falls out.
 
-That is what a **shader** is, and it is why an effect written this way is
-resolution-independent: the same code fills a 16-light strip and a 12,288-light
-wall, because it never counted pixels in the first place.
+That is what a **shader** is, and it is why an effect written this way is resolution-independent: the same code fills a 16-light strip and a 12,288-light wall, because it never counted pixels in the first place.
 
-Fields also compose. Add two, and you get a third. Use one to bend another's
-coordinates, and you get something neither could produce alone. That is the whole
-trick, and the rest of this page is the vocabulary.
+Fields also compose. Add two, and you get a third. Use one to bend another's coordinates, and you get something neither could produce alone. That is the whole trick, and the rest of this page is the vocabulary.
 
 ---
 
 ## 2. Noise: the raw material
 
-Random numbers look like static. **Noise** looks like nature, because a noise
-value is close to its neighbors: it wanders instead of jumping.
+Random numbers look like static. **Noise** looks like nature, because a noise value is close to its neighbors: it wanders instead of jumping.
 
 **Ken Perlin** published it in 1985 (*An Image Synthesizer*), after developing it
-on the movie *Tron*; it later won him an Academy Award. projectMM uses his own
-revision, *improved noise* (SIGGRAPH 2002), which removes the directional bias of
-the original.
+on the movie *Tron*; it later won him an Academy Award. projectMM uses his own revision, *improved noise* (SIGGRAPH 2002), which removes the directional bias of the original.
 
 ```
 int v = noise(x * 20, y * 20, div(t, 32));   // 0..255, smooth in every direction
@@ -52,31 +37,24 @@ Three things to know, and then you can use it:
 
 **Scale is everything.** `noise(x, y, 0)` reads one value per light and looks like
 static, because adjacent lights land in completely different parts of the field.
-Multiply the coordinate down and you zoom in: `noise(x * 20, y * 20, 0)` gives
-broad soft blobs. Getting this wrong is the single most common reason a noise
-effect looks like colored snow.
+Multiply the coordinate down and you zoom in: `noise(x * 20, y * 20, 0)` gives broad soft blobs. Getting this wrong is the single most common reason a noise effect looks like colored snow.
 
 **Time is just another axis.** Do not recompute a 2D field and shift it. Sample a
-3D field and walk along z: `noise(x * 20, y * 20, div(t, 32))`. The picture morphs
-instead of scrolling, which is what makes it look organic rather than mechanical.
+3D field and walk along z: `noise(x * 20, y * 20, div(t, 32))`. The picture morphs instead of scrolling, which is what makes it look organic rather than mechanical.
 
 **1D, 2D and 3D are the same call.** One axis for a strip, two for a panel, three
 for a cube, or two of space and one of time. `noise()` takes what you give it.
 
-(`div(t, 32)` rather than `t / 32`: MoonLive spells integer division `div`, and
-`t` is the millisecond clock every script is handed.)
+(`div(t, 32)` rather than `t / 32`: MoonLive spells integer division `div`, and `t` is the millisecond clock every script is handed.)
 
 ---
 
 ## 3. Octaves: detail at every size
 
-One noise sample is a soft blur. Nature is not soft: a coastline has bays, and the
-bays have inlets, and the inlets have rocks.
+One noise sample is a soft blur. Nature is not soft: a coastline has bays, and the bays have inlets, and the inlets have rocks.
 
 **fBm** (fractional Brownian motion, the term is Mandelbrot's) is the answer, and
-it is embarrassingly simple. Sample the field several times, each at double the
-frequency and half the strength, and add them up. Each pass is an **octave**, the
-name borrowed from music, where an octave is also a doubling.
+it is embarrassingly simple. Sample the field several times, each at double the frequency and half the strength, and add them up. Each pass is an **octave**, the name borrowed from music, where an octave is also a doubling.
 
 ```
 int v = fbm(x * 20, y * 20, 2);   // 2 octaves: shape, plus texture on it
@@ -88,78 +66,51 @@ int v = fbm(x * 20, y * 20, 2);   // 2 octaves: shape, plus texture on it
 | 2 | shape with texture on it | 2 samples |
 | 4 | rock, cloud, terrain | 4 samples |
 
-Octaves are the **cost knob** of every field effect: doubling them doubles the
-work per light. Two is usually the sweet spot; four is for when the fixture is
-small enough to afford it.
+Octaves are the **cost knob** of every field effect: doubling them doubles the work per light. Two is usually the sweet spot; four is for when the fixture is small enough to afford it.
 
-One measured detail worth knowing: octaves partly cancel, so their sum is
-narrower than one octave's range. Four octaves span roughly 54..199 of 0..255
-rather than the full sweep. projectMM re-widens the result for you, so `fbm` at
-any octave count still uses the whole range and your thresholds keep working.
+One measured detail worth knowing: octaves partly cancel, so their sum is narrower than one octave's range. Four octaves span roughly 54..199 of 0..255 rather than the full sweep. projectMM re-widens the result for you, so `fbm` at any octave count still uses the whole range and your thresholds keep working.
 
 ---
 
 ## 4. Warp: the trick that sells it
 
-This one is worth learning on its own, because it is the difference between "some
-noise" and "that looks like smoke".
+This one is worth learning on its own, because it is the difference between "some noise" and "that looks like smoke".
 
 **Domain warping** (popularized by **Inigo Quilez**, of Shadertoy) means using
-noise to move the place where you sample noise. You are not adding a field to a
-field, you are bending one field's coordinates with another.
+noise to move the place where you sample noise. You are not adding a field to a field, you are bending one field's coordinates with another.
 
 ```
 int v = warp(x * 20, y * 20, 60);   // 60 = how far the field displaces itself
 ```
 
-The result flows, folds and curls back on itself. Almost every "liquid" or
-"marbled" look you have admired is domain warping.
+The result flows, folds and curls back on itself. Almost every "liquid" or "marbled" look you have admired is domain warping.
 
 ![Aurora: layered warped noise read in polar coordinates](../assets/light/tutorial/aurora.png)
 
-That is **Aurora**: a few layers of warped noise, each drifting on its own clock,
-read around the center instead of across the grid, with a contrast window that
-pushes most of the field to black so distinct curtains survive.
+That is **Aurora**: a few layers of warped noise, each drifting on its own clock, read around the center instead of across the grid, with a contrast window that pushes most of the field to black so distinct curtains survive.
 
-Every technique on this page is a published one, credited to the people named
-throughout. **Stefan Petrick** is the person who brought that shader vocabulary to
-LED panels and showed what it does there, through
-[Animartrix](https://github.com/StefanPetrick/animartrix), FunkyNoise and
-ColorTrails: noise read in polar coordinates, layers on independent oscillators, a
-contrast window that turns a soft field into distinct curtains, emitters carried by
-a flow field. That is the tradition this page's effects sit in.
+Every technique on this page is a published one, credited to the people named throughout. **Stefan Petrick** is the person who brought that shader vocabulary to LED panels and showed what it does there, through [Animartrix](https://github.com/StefanPetrick/animartrix), FunkyNoise and ColorTrails: noise read in polar coordinates, layers on independent oscillators, a contrast window that turns a soft field into distinct curtains, emitters carried by a flow field. That is the tradition this page's effects sit in.
 
 ---
 
 ## 5. From lines to fields: the ladder
 
-The power functions are a ladder, and the rungs are worth knowing because each one
-buys a different kind of beauty.
+The power functions are a ladder, and the rungs are worth knowing because each one buys a different kind of beauty.
 
 **Lines and shapes** are the bottom rung. `line` (Bresenham 1962), `circle`
-(Bresenham's 1965 midpoint form), `disc`, `sphere`, `text`, `sprite`. Exact,
-cheap, and the right answer when you want a *thing* on the panel rather than a
-texture.
+(Bresenham's 1965 midpoint form), `disc`, `sphere`, `text`, `sprite`. Exact, cheap, and the right answer when you want a *thing* on the panel rather than a texture.
 
 **Signed distance fields** are the same shapes, made soft. Instead of "is this
-pixel inside the circle", `sdCircle` answers "how far is this pixel from the
-edge", negative inside. A distance is a number you can threshold, glow, outline or
-blend, and two distance fields blend into shapes neither one describes. The
-catalog follows **Inigo Quilez**'s.
+pixel inside the circle", `sdCircle` answers "how far is this pixel from the edge", negative inside. A distance is a number you can threshold, glow, outline or blend, and two distance fields blend into shapes neither one describes. The catalog follows **Inigo Quilez**'s.
 
 **Particles** give you motion with memory. A particle pool carries position and
-velocity, and `gravity`, `drag`, `bounce` and `collide` act on all of them at
-once. Use this when the thing that matters is that a spark *fell*, and the pixels
-are just where it happens to be.
+velocity, and `gravity`, `drag`, `bounce` and `collide` act on all of them at once. Use this when the thing that matters is that a spark *fell*, and the pixels are just where it happens to be.
 
 **Fields** are where this page started: noise, fBm, warp, sampled per light.
 
 **Transport** is the top rung, and it is what this release added. A field says
 where things go; transport actually *carries* light along it, frame after frame.
-`advect` moves a whole plane along a velocity field by asking, for each
-destination, where its contents came from. `decay` fades what is already there by
-a half-life in milliseconds. Together they make a trail that is not drawn: it is
-the previous frames' light, moved and dimmed.
+`advect` moves a whole plane along a velocity field by asking, for each destination, where its contents came from. `decay` fades what is already there by a half-life in milliseconds. Together they make a trail that is not drawn: it is the previous frames' light, moved and dimmed.
 
 **Curl noise** (**Robert Bridson**, 2007) is the field transport wants. Take a
 noise field and use its perpendicular gradient, and the result is
@@ -169,9 +120,7 @@ property is why a curl flow reads as a real medium rather than as arrows.
 ![Nebula: a noise field born into a curl flow](../assets/light/tutorial/nebula.png)
 
 **Simulation** is the last rung. **Jos Stam**'s *Stable Fluids* (SIGGRAPH 1999)
-solves the medium's own motion: diffuse, project, advect, project. Every other
-flow is a function of position and time; this one is state. A jet fired now
-changes where everything downstream goes for seconds afterwards.
+solves the medium's own motion: diffuse, project, advect, project. Every other flow is a function of position and time; this one is state. A jet fired now changes where everything downstream goes for seconds afterwards.
 
 ![Fluid: dye poured into a Stam solver](../assets/light/tutorial/fluid.png)
 
@@ -179,35 +128,20 @@ changes where everything downstream goes for seconds afterwards.
 
 ## 6. Why 16 bits, in one paragraph
 
-A trail is a value multiplied by slightly less than one, sixty times a second. At
-8 bits that does not work: a value of 100 multiplied by 0.994 either truncates
-back to 99 and keeps falling too fast, or rounds back to 100 and never fades at
-all. The tail either vanishes or freezes.
+A trail is a value multiplied by slightly less than one, sixty times a second. At 8 bits that does not work: a value of 100 multiplied by 0.994 either truncates back to 99 and keeps falling too fast, or rounds back to 100 and never fades at all. The tail either vanishes or freezes.
 
-So every effect that carries light over time keeps its own plane at **16 bits**
-and narrows to 8 once, at the end, with dithering. The layer buffer stays 8-bit,
-which keeps every driver fast. You get this for free: `trail(1)` in a script
-allocates it, and the blit happens for you.
+So every effect that carries light over time keeps its own plane at **16 bits** and narrows to 8 once, at the end, with dithering. The layer buffer stays 8-bit, which keeps every driver fast. You get this for free: `trail(1)` in a script allocates it, and the blit happens for you.
 
 ---
 
 ## 7. Compiled or scripted
 
-Much of the vocabulary above exists twice: as a C++ kernel, and as a MoonLive builtin. Same
-algorithm, same numbers. The shapes are worth knowing, though, because they are not identical.
+Much of the vocabulary above exists twice: as a C++ kernel, and as a MoonLive builtin. Same algorithm, same numbers. The shapes are worth knowing, though, because they are not identical.
 
-The noise family (`noise`, `fbm`, `warp`), `line`, `circle` and the palette calls are the same
-function under both names. The **transport** kernels are reached differently: `advect16`, `decay16`,
-`quantize`, `blit16`, `upscale16` and `halfLifeKeep` are compiled-only, and a script uses them
-through six builtins that wrap them, `trail(1)`, `flowNoise`, `flowCurl`, `trailDecay`, `emitTrail`
-and `fieldRate`. That is deliberate: a script names a wind and a persistence, and the binding owns
-the planes, the ping-pong and the narrowing, which is what keeps a whole-plane operation one host
-call rather than one per light. Compiled-only for now: `disc`, `sphere`, `text`, `sprite`, the SDF
-catalog and the fluid solver.
+The noise family (`noise`, `fbm`, `warp`), `line`, `circle` and the palette calls are the same function under both names. The **transport** kernels are reached differently: `advect16`, `decay16`, `quantize`, `blit16`, `upscale16` and `halfLifeKeep` are compiled-only, and a script uses them through six builtins that wrap them, `trail(1)`, `flowNoise`, `flowCurl`, `trailDecay`, `emitTrail` and `fieldRate`. That is deliberate: a script names a wind and a persistence, and the binding owns the planes, the ping-pong and the narrowing, which is what keeps a whole-plane operation one host call rather than one per light. Compiled-only for now: `disc`, `sphere`, `text`, `sprite`, the SDF catalog and the fluid solver.
 
 **MoonLive** is projectMM's scripting language. Scripts are compiled to native
-code on the device, so a script is not interpreted per pixel: it runs at machine
-speed. You edit one in the browser and the picture changes as you type.
+code on the device, so a script is not interpreted per pixel: it runs at machine speed. You edit one in the browser and the picture changes as you type.
 
 The interesting number is how little a script has to say:
 
@@ -218,21 +152,11 @@ The interesting number is how little a script has to say:
 | Aurora | 58 lines | 216 lines |
 | Nebula | 63 lines | 271 lines |
 
-These are not stripped-down versions: each is a real effect someone would happily
-run. The reason they are a fifth of the size is that **the heavy lifting is in the
-kernels, and the kernels are available to both.** A script does not implement
-advection; it calls it. The C++ version is longer mostly because it also handles
-allocation, resizing and control registration, which the script binding does on the
-script's behalf.
+These are not stripped-down versions: each is a real effect someone would happily run. The reason they are a fifth of the size is that **the heavy lifting is in the kernels, and the kernels are available to both.** A script does not implement advection; it calls it. The C++ version is longer mostly because it also handles allocation, resizing and control registration, which the script binding does on the script's behalf.
 
-One honest caveat on that table: `fluid.mle` is not the Stam solver. The solver has no script
-binding yet, so the script pours jets into a curl flow, which LOOKS like a fluid without simulating
-one. Aurora, Trails and Nebula are genuinely the same effect in both columns.
+One honest caveat on that table: `fluid.mle` is not the Stam solver. The solver has no script binding yet, so the script pours jets into a curl flow, which LOOKS like a fluid without simulating one. Aurora, Trails and Nebula are genuinely the same effect in both columns.
 
-This is the point where the two halves reinforce each other. Every kernel added
-for a compiled effect immediately makes scripts more capable, and every effect
-that turns out to be expressible as a script is one that does not need to be
-compiled in at all. `fluid.mle` is 44 lines:
+This is the point where the two halves reinforce each other. Every kernel added for a compiled effect immediately makes scripts more capable, and every effect that turns out to be expressible as a script is one that does not need to be compiled in at all. `fluid.mle` is 44 lines:
 
 ```
 void tick() {
@@ -250,34 +174,22 @@ void tick() {
 Three kernel calls and a loop. That is a fluid-looking effect.
 
 **Which to use.** Script first, always: you iterate in seconds instead of minutes,
-and a script that fails to compile leaves the previous one running rather than
-taking the device down. It is bounded rather than sandboxed, though. A script gets
-a fixed member budget and a fixed number of live variables, and the compiler
-refuses a script that exceeds them; what it cannot refuse is a per-light loop too
-expensive for the fixture, which shows up as a slow frame rate and, on a large
-grid, can starve the network stack enough that the web UI stops answering while
-the lights keep running. Move to C++ when you need something the vocabulary does
-not have, or when the per-light cost is the problem (`fieldRate(n)` exists
-precisely to make a per-light loop affordable on a big fixture).
+and a script that fails to compile leaves the previous one running rather than taking the device down. It is bounded rather than sandboxed, though. A script gets a fixed member budget and a fixed number of live variables, and the compiler refuses a script that exceeds them; what it cannot refuse is a per-light loop too expensive for the fixture, which shows up as a slow frame rate and, on a large grid, can starve the network stack enough that the web UI stops answering while the lights keep running. Move to C++ when you need something the vocabulary does not have, or when the per-light cost is the problem (`fieldRate(n)` exists precisely to make a per-light loop affordable on a big fixture).
 
 ---
 
 ## 8. Everything is 3D, when the fixture is
 
 Every kernel on this page takes a depth. On a panel you pass 1 and pay nothing:
-the loop is the 2D loop exactly. On a cube, the field is sampled along a real
-third axis, so the slices differ instead of one plane repeating.
+the loop is the 2D loop exactly. On a cube, the field is sampled along a real third axis, so the slices differ instead of one plane repeating.
 
-Worth knowing where the line currently is: the **fields** are genuinely
-volumetric, while **transport** carries light within each slice and not yet
-between them. A trail on a cube travels across its slice, not through the volume.
+Worth knowing where the line currently is: the **fields** are genuinely volumetric, while **transport** carries light within each slice and not yet between them. A trail on a cube travels across its slice, not through the volume.
 
 ---
 
 ## 9. Try this
 
-The fastest way to feel it is to break something on purpose. Open a MoonLive
-script, pick one number, and move it a long way.
+The fastest way to feel it is to break something on purpose. Open a MoonLive script, pick one number, and move it a long way.
 
 | Change | What you should see | The lesson |
 |---|---|---|
@@ -291,20 +203,14 @@ script, pick one number, and move it a long way.
 
 ## 10. Making one good: what the failures teach
 
-The kernels above are the easy half. An effect built correctly on them can still be
-dull, and eight built in one day produced four keepers: the other four were dropped
-for being unattractive, too sparse, too slow, or for dying out while the music played.
+The kernels above are the easy half. An effect built correctly on them can still be dull, and eight built in one day produced four keepers: the other four were dropped for being unattractive, too sparse, too slow, or for dying out while the music played.
 What follows is what those four cost, written down so the next effect skips them.
 
 ### Measure the picture, do not theorize about it
 
-Every wrong diagnosis on this page came from reasoning about code instead of reading a
-number off the frame. Rendering to a buffer and counting is cheap: what fraction of
-lights are lit, what the brightest one is, how far the picture moves between frames.
+Every wrong diagnosis on this page came from reasoning about code instead of reading a number off the frame. Rendering to a buffer and counting is cheap: what fraction of lights are lit, what the brightest one is, how far the picture moves between frames.
 
-One effect rendered pure black for a day. The cause was a scale mismatch (a wave pressed
-15 units deep while the renderer divided slopes by 512, so every ripple came out below the
-visibility threshold), and no amount of reading the code found it. One measurement did:
+One effect rendered pure black for a day. The cause was a scale mismatch (a wave pressed 15 units deep while the renderer divided slopes by 512, so every ripple came out below the visibility threshold), and no amount of reading the code found it. One measurement did:
 brightest light 0, in a simulation whose physics was fine. Cheap checks, in order:
 
 | Reads | Means |
@@ -314,76 +220,46 @@ brightest light 0, in a simulation whose physics was fine. Cheap checks, in orde
 | brightest light far below 255 | a scale mismatch between what is computed and what is displayed |
 | frames identical over time | the simulation has converged, or the clock is not advancing |
 
-Pick a quantity that cannot cancel. A center of mass on a symmetric flow stays put while
-every parcel moves, and reports "nothing is happening" about a working effect.
+Pick a quantity that cannot cancel. A center of mass on a symmetric flow stays put while every parcel moves, and reports "nothing is happening" about a working effect.
 
 ### Match the scales at every seam
 
-Half of the day's bugs were one number expressed in the wrong unit. A control is 0..255,
-a height field rings to ±20000, a fixed-point position carries 16 fractional bits. Every
-place two of those meet is a place where a plausible-looking line silently produces
-nothing, and nothing about it looks wrong on the screen.
+Half of the day's bugs were one number expressed in the wrong unit. A control is 0..255, a height field rings to ±20000, a fixed-point position carries 16 fractional bits. Every place two of those meet is a place where a plausible-looking line silently produces nothing, and nothing about it looks wrong on the screen.
 
-Write the conversion where the value crosses, name the constant, and say in a comment what
-range each side speaks in. When something renders black or blindingly white, suspect a
-seam before suspecting the algorithm.
+Write the conversion where the value crosses, name the constant, and say in a comment what range each side speaks in. When something renders black or blindingly white, suspect a seam before suspecting the algorithm.
 
 ### A simulation converges: keep feeding it
 
-A fluid driven by forces at fixed positions and fixed angles reaches a steady state and
-stops. The picture goes still while the music keeps playing, which reads as a crash. This
-killed one effect outright.
+A fluid driven by forces at fixed positions and fixed angles reaches a steady state and stops. The picture goes still while the music keeps playing, which reads as a crash. This killed one effect outright.
 
-Anything that integrates its own state needs its input to keep changing: rotate where the
-forces are applied, lean their direction, alternate their sense. Motion in the input is
-what keeps motion in the output.
+Anything that integrates its own state needs its input to keep changing: rotate where the forces are applied, lean their direction, alternate their sense. Motion in the input is what keeps motion in the output.
 
 ### Frame-rate independence is not optional
 
-A simulation stepped once per rendered frame runs at whatever speed the hardware happens
-to deliver. Measured here: 16x faster on a desktop at 1200 fps than on a device at 60. The
-same effect that reads well on a panel is unusable on a bench.
+A simulation stepped once per rendered frame runs at whatever speed the hardware happens to deliver. Measured here: 16x faster on a desktop at 1200 fps than on a device at 60. The same effect that reads well on a panel is unusable on a bench.
 
-Accumulate elapsed time and step a fixed amount (16 ms works), capping the catch-up so a
-stall cannot spend a second of frames at once. Everything that moves goes inside that loop,
-including trail fades: leaving the fade outside makes trail length depend on frame rate
-even when the motion does not.
+Accumulate elapsed time and step a fixed amount (16 ms works), capping the catch-up so a stall cannot spend a second of frames at once. Everything that moves goes inside that loop, including trail fades: leaving the fade outside makes trail length depend on frame rate even when the motion does not.
 
 ### Fill the screen, and keep filling it
 
-Two effects were dropped for the same reason: at rest they showed too little. An effect is
-judged on the whole panel, so a good one covers it, and covers it in the first second.
+Two effects were dropped for the same reason: at rest they showed too little. An effect is judged on the whole panel, so a good one covers it, and covers it in the first second.
 Sparse output at 6% of lights is a demo, not an effect.
 
-Two habits fix most of it. Give the effect something to do with no input at all, so it is
-never blank while someone waits for a beat. And start its clocks primed rather than at
-zero: an effect whose first event is two seconds away reads as broken long before it reads
-as calm.
+Two habits fix most of it. Give the effect something to do with no input at all, so it is never blank while someone waits for a beat. And start its clocks primed rather than at zero: an effect whose first event is two seconds away reads as broken long before it reads as calm.
 
 ### Slow is a design property, not a tuning problem
 
-An effect that misses its frame time on the target device is not a slow effect to be
-optimized later. On the S31 one solver-per-frame effect was both too slow and too dull,
-which is the common case rather than a coincidence: the expensive part was not the part
-doing the visual work. Decide what the effect spends its budget on before building it, and
-measure on the smallest device it claims to support.
+An effect that misses its frame time on the target device is not a slow effect to be optimized later. On the S31 one solver-per-frame effect was both too slow and too dull, which is the common case rather than a coincidence: the expensive part was not the part doing the visual work. Decide what the effect spends its budget on before building it, and measure on the smallest device it claims to support.
 
 ### Two failed attempts means stop
 
-The rule the project already carries applies hardest here, because a picture always
-suggests one more plausible tweak. Two attempts that do not fix it mean the diagnosis is
-wrong, not the parameters. On this page a second "fix" made an effect visibly worse than
-the bug it targeted.
+The rule the project already carries applies hardest here, because a picture always suggests one more plausible tweak. Two attempts that do not fix it mean the diagnosis is wrong, not the parameters. On this page a second "fix" made an effect visibly worse than the bug it targeted.
 
 ### Pin the behavior, not just the pixels
 
-A golden-hash test passes happily on an all-black frame, which is how an effect shipped
-rendering nothing while its test stayed green. A hash pins *which* pixels light; it says
-nothing about whether any of them do.
+A golden-hash test passes happily on an all-black frame, which is how an effect shipped rendering nothing while its test stayed green. A hash pins *which* pixels light; it says nothing about whether any of them do.
 
-Pin the property that would have caught the failure: that the surface is visible, that it
-still moves after a hundred frames. Then confirm the test fails against the broken code
-before trusting it, because a test that passes on both is measuring nothing.
+Pin the property that would have caught the failure: that the surface is visible, that it still moves after a hundred frames. Then confirm the test fails against the broken code before trusting it, because a test that passes on both is measuring nothing.
 
 ## Where to go next
 
diff --git a/docs/tutorials/how-projectmm-works.md b/docs/tutorials/how-projectmm-works.md
index c6bc776e..e93e9d1e 100644
--- a/docs/tutorials/how-projectmm-works.md
+++ b/docs/tutorials/how-projectmm-works.md
@@ -2,9 +2,7 @@
 
 You have lights running. This page explains what you were actually looking at.
 
-There is really only **one idea** in projectMM, and everything else follows from
-it. Ten minutes here and the rest of the interface stops being a wall of settings
-and becomes a place you know your way around.
+There is really only **one idea** in projectMM, and everything else follows from it. Ten minutes here and the rest of the interface stops being a wall of settings and becomes a place you know your way around.
 
 > New here? Start with **[Install & first light](../gettingstarted.md)** and come
 > back once something is lit.
@@ -17,11 +15,9 @@ and becomes a place you know your way around.
 
 ## 1. Everything is a card
 
-Every single thing in projectMM — the WiFi settings, an effect, an LED driver,
-the file manager — is a **MoonModule**. And every MoonModule is a **card**.
+Every single thing in projectMM, the WiFi settings, an effect, an LED driver, the file manager, is a **MoonModule**. And every MoonModule is a **card**.
 
-There is no second kind of thing. Learn to read one card and you can read them
-all.
+There is no second kind of thing. Learn to read one card and you can read them all.
 
 ![A Layouts card with a Grid card nested inside it](../assets/gettingstarted/02-08-UI-Layouts.png)
 
@@ -36,31 +32,26 @@ Look at that picture again, because every card has the same parts:
 | **?** | Open this module's documentation |
 | **`{ }`** | Open its live state, for [issue reports](../how-to/logging-an-issue.md) |
 | `status`, `width`, `height` … | The settings, one per row |
-| The **Grid** box inside | A child card — cards nest |
+| The **Grid** box inside | A child card, cards nest |
 | **+ add module** | Add a child here |
 
 Two of those are unusual and worth pausing on.
 
 **The timing and memory are per module, and they are live.** Most software hides
-this. projectMM shows you exactly what each part costs, so when something feels
-slow you can see which card is responsible instead of guessing.
+this. projectMM shows you exactly what each part costs, so when something feels slow you can see which card is responsible instead of guessing.
 
 **The nesting is the structure.** A child card is drawn *inside* its parent's
-border. When you see `Grid` inside `Layouts`, that is not decoration — the Grid
-really does belong to Layouts. The box is the truth.
+border. When you see `Grid` inside `Layouts`, that is not decoration, the Grid really does belong to Layouts. The box is the truth.
 
 ---
 
 ## 2. Two families of card
 
-Cards come in two kinds, and knowing which is which tells you where to look for
-anything.
+Cards come in two kinds, and knowing which is which tells you where to look for anything.
 
-### Core — the device itself
+### Core: the device itself
 
-Identity, network, files, firmware. This half knows *nothing* about lights: it
-would be identical on a machine driving motors or synthesising audio. These cards
-are fixed — you configure them, you don't add or remove them.
+Identity, network, files, firmware. This half knows *nothing* about lights: it would be identical on a machine driving motors or synthesising audio. These cards are fixed, you configure them, you don't add or remove them.
 
 ![The System card, showing device name, chip, memory and uptime](../assets/gettingstarted/02-05-UI-System.png)
 
@@ -70,57 +61,48 @@ are fixed — you configure them, you don't add or remove them.
 | **Network** | WiFi and Ethernet, plus MQTT and finding other devices |
 | **File Manager** | Browse, edit and upload files on the device |
 | **Firmware** | Your version, and updates over the air |
-| **Services** | Hardware that isn't lighting — a microphone, a sensor |
+| **Services** | Hardware that isn't lighting, a microphone, a sensor |
 | **Control** | Drive controls from MIDI, a remote, or another device |
 
-### Light — what you build with
+### Light: what you build with
 
-Lights, colours, and getting the bytes out. These cards are yours: add them,
-remove them, rearrange them freely.
+Lights, colors, and getting the bytes out. These cards are yours: add them, remove them, rearrange them freely.
 
 | Card | What it's for |
 |---|---|
 | **Layouts** | Where the lights physically ARE |
 | **Effects** | What color they are, over time |
-| **Drivers** | How the colours actually reach the lights |
+| **Drivers** | How the colors actually reach the lights |
 
-That split is why the menu is ordered the way it is: the device first, then the
-three light cards.
+That split is why the menu is ordered the way it is: the device first, then the three light cards.
 
 ---
 
 ## 3. The light pipeline
 
-Those three light cards are not a random list. They are a **pipeline**, and they
-run in order, every frame:
+Those three light cards are not a random list. They are a **pipeline**, and they run in order, every frame:
 
 ```text
    Layouts    ──→    Effects    ──→    Drivers
-    where            what colour        how it gets out
+    where            what color        how it gets out
 ```
 
 <!-- IMAGE PLACEHOLDER: a friendly left-to-right diagram of the three stages.
-     Stage 1 shows a grid of grey dots (positions only); stage 2 the same grid
-     with colour flowing across it; stage 3 an arrow leaving toward a real LED
-     strip and a network cable. This is the single most valuable new image on
-     the page — the pipeline is the concept everything else hangs off. -->
+     Stage 1 shows a grid of grey dots (positions only); stage 2 the same grid with color flowing across it; stage 3 an arrow leaving toward a real LED strip and a network cable. This is the single most valuable new image on the page, the pipeline is the concept everything else hangs off. -->
 
 **1. A Layout decides where the lights are.**
-A 16×16 grid, a ring of 60, two rows of 24, a shape you draw yourself. It answers
-one question: *how many lights, and where is each one?*
+A 16×16 grid, a ring of 60, two rows of 24, a shape you draw yourself. It answers one question: *how many lights, and where is each one?*
 
 ![Layouts](../assets/light/Layouts.png)
 
 **2. An Effect paints them.**
-It asks the layout for the coordinates and writes colours — every frame, forever.
-Effects live inside a **Layer**, and a Layer can hold several, stacked and blended
-like layers in an image editor.
+It asks the layout for the coordinates and writes colors, every frame, forever.
+Effects live inside a **Layer**, and a Layer can hold several, stacked and blended like layers in an image editor.
 
 ![A plasma effect running](../assets/light/effects/PlasmaEffect.gif)
 
 **3. A Driver sends the result somewhere.**
-Out a GPIO pin to a strip, across the network as ArtNet or DDP, or to the preview
-in your browser.
+Out a GPIO pin to a strip, across the network as ArtNet or DDP, or to the preview in your browser.
 
 ![Drivers](../assets/light/Drivers.png)
 
@@ -128,18 +110,13 @@ in your browser.
 
 Because the stages are separate, **changing one doesn't break the others**.
 
-Resize your panel from 16×16 to 32×32 and every effect keeps running — it simply
-asks the layout a new question and gets new coordinates back. That is why the
-same effect works on a grid, a ring, or a wall of 12,288 lights without being
-rewritten.
+Resize your panel from 16×16 to 32×32 and every effect keeps running, it simply asks the layout a new question and gets new coordinates back. That is why the same effect works on a grid, a ring, or a wall of 12,288 lights without being rewritten.
 
-**Modifiers** sit between layout and effect and bend the mapping — mirror it,
-fold it, tile it. An effect written for one strip can drive a whole wall without
-ever knowing it.
+**Modifiers** sit between layout and effect and bend the mapping, mirror it,
+fold it, tile it. An effect written for one strip can drive a whole wall without ever knowing it.
 
 <!-- IMAGE PLACEHOLDER: before/after of one effect with a mirror modifier
-     applied — the same frame, un-mirrored and mirrored, side by side. Makes
-     "a modifier folds the mapping" obvious in one glance. -->
+     applied, the same frame, un-mirrored and mirrored, side by side. Makes "a modifier folds the mapping" obvious in one glance. -->
 
 ---
 
@@ -147,19 +124,14 @@ ever knowing it.
 
 **Settings need no save button, and nothing needs a reboot.**
 
-Move a slider and the next frame uses it. Change the WiFi credentials, add an
-effect, resize a grid — all of it takes effect immediately, on a running device.
+Move a slider and the next frame uses it. Change the WiFi credentials, add an effect, resize a grid, all of it takes effect immediately, on a running device.
 
 <video src="../assets/uiscenarios/swap-an-effect.webm" autoplay loop muted playsinline width="720" title="Replacing an effect with another type; the slot keeps its place and the lights never stop"></video>
 
 **Files are the one exception, and only because typing is different.** A script
-you are editing is saved when you click away, press Ctrl/Cmd+S, or press Save —
-a half-typed line should not be compiled onto your fixture mid-word. The moment
-it is saved it recompiles and swaps in live, same as everything else.
+you are editing is saved when you click away, press Ctrl/Cmd+S, or press Save, a half-typed line should not be compiled onto your fixture mid-word. The moment it is saved it recompiles and swaps in live, same as everything else.
 
-This is a deliberate design rule, not a convenience, and you can trust it: if
-something needs a reboot before it works, that is a bug worth
-[reporting](../how-to/logging-an-issue.md).
+This is a deliberate design rule, not a convenience, and you can trust it: if something needs a reboot before it works, that is a bug worth [reporting](../how-to/logging-an-issue.md).
 
 Your settings save themselves and survive a power cycle.
 
@@ -167,10 +139,7 @@ Your settings save themselves and survive a power cycle.
 
 ## 5. The emoji on every card
 
-Each card and every row of the picker carries a few emoji. They are a filter, not
-decoration: the picker's chip row lets you narrow a long list to the effects that
-listen to music, or the layouts that build a volume. So an emoji only exists where
-it groups several modules, and a module that fits no group carries none.
+Each card and every row of the picker carries a few emoji. They are a filter, not decoration: the picker's chip row lets you narrow a long list to the effects that listen to music, or the layouts that build a volume. So an emoji only exists where it groups several modules, and a module that fits no group carries none.
 
 Three come from what the module IS, and the interface adds them for you:
 
@@ -191,9 +160,7 @@ The rest the module declares about itself:
 | 👾 | pixel art: the games and sprites |
 | 🧬 | a simulation: the picture emerges from cells evolving off their own last frame, rather than being drawn |
 
-And a last group, shown together at the end of the row, saying which **power functions** the effect
-is built on. These are the kernels of [the shared library](../moonmodules/light/power-functions.md),
-so they group effects by what they are made of, and by what that makes them look like:
+And a last group, shown together at the end of the row, saying which **power functions** the effect is built on. These are the kernels of [the shared library](../moonmodules/light/power-functions.md), so they group effects by what they are made of, and by what that makes them look like:
 
 | | |
 |---|---|
@@ -213,9 +180,9 @@ A module can carry several: `💫🎶` is a MoonLight effect that reacts to freq
 
 Now that you know what the cards are, go build something:
 
-- **[Effects](../moonmodules/light/effects.md)** — the catalog, with pictures
-- **[Layouts](../moonmodules/light/layouts.md)** — grids, rings and custom shapes
-- **[Modifiers](../moonmodules/light/modifiers.md)** — mirroring, folding, tiling
-- **[Drivers](../moonmodules/light/drivers.md)** — LED strips, ArtNet, DMX
+- **[Effects](../moonmodules/light/effects.md)**: the catalog, with pictures
+- **[Layouts](../moonmodules/light/layouts.md)**: grids, rings and custom shapes
+- **[Modifiers](../moonmodules/light/modifiers.md)**: mirroring, folding, tiling
+- **[Drivers](../moonmodules/light/drivers.md)**: LED strips, ArtNet, DMX
 - **[Live scripting](../moonmodules/light/moonlive.md)**: write your own
   effect in the browser; it compiles on the device and runs as native code
diff --git a/moondeck/MoonDeck.md b/moondeck/MoonDeck.md
index d3f78c90..9cc4aad9 100644
--- a/moondeck/MoonDeck.md
+++ b/moondeck/MoonDeck.md
@@ -182,7 +182,7 @@ And per file, over every `.h` and `.cpp` under `src/`, `test/`, `esp32/main`, `m
 
 **A finding in a file that generates a page is an error; one in an implementation file is a warning.** A header and a catalog page are published, so a defect there ships and fails the gate. A `.cpp` publishes nothing: its comments are a note to the next reader, worth fixing without stopping a commit. Both are counted and both are reported. The split stages the sweep rather than ranking the two kinds of comment, so it goes and everything blocks once the warning column reaches zero.
 
-Every run writes [docs/reference/metrics/docgen.md](../docs/reference/metrics/docgen.md), the tracked state of the sweep: errors and warnings per rule and per page, then the files ranked within each area. Current state only, so its git history is the trend, the same shape repo-health.md uses. The file is the artifact to read, because stdout scrolls away and truncates; `--noreport` suppresses the write for a caller that only wants the exit code.
+Every run writes [docs/reference/metrics/docgen.md](../docs/reference/metrics/docgen.md), the tracked state of the sweep: errors and warnings per rule and per page, then the files ranked within each area. Current state only, so its git history is the trend, the same shape repo-health.md uses. Every erroring file also carries its own findings in full, since an error blocks a commit and its reader needs to know what to fix; warnings stay counted rather than listed, which keeps the page a report rather than a log. The file is the artifact to read, because stdout scrolls away and truncates.
 
 **A ratchet, not a snapshot.** The committed `docs/reference/metrics/docgen.md` is the baseline: the run rewrites the file, then compares what it found against the copy in `HEAD`. An error fails the run, and so does any rule whose warning count ROSE, the total included. Warnings are staged work, and staged work that grows is not a sweep.
 
diff --git a/moondeck/check/check_devices.py b/moondeck/check/check_devices.py
index da8e6af6..02316807 100644
--- a/moondeck/check/check_devices.py
+++ b/moondeck/check/check_devices.py
@@ -92,7 +92,10 @@ def eth_preset_drift():
 
     # label -> the PHY constant naming that chip's arm of the ethConfigDefault ternary.
     PAIRS = {"Classic RMII": "ethLan8720", "P4-NANO": "ethIp101", "S31 CoreBoard": "ethYt8531"}
-    FIELDS = ["phyAddr", "mdc", "mdio", "rst", "rmiiClk"]
+    # Every field `seedEthPresetFromPins` compares, or a drift it ignores reseeds a board to Custom.
+    FIELDS = ["phyAddr", "mdc", "mdio", "rst", "rmiiClk", "rmiiClockExtIn"]
+    # The preset table writes the PHY as its enum VALUE and the header names the constant.
+    PHY_VALUE = {"ethLan8720": "1", "ethIp101": "2", "ethW5500": "3", "ethYt8531": "4"}
     drift = []
     for label, phy in PAIRS.items():
         row = rows.get(label)
@@ -100,11 +103,14 @@ def eth_preset_drift():
             drift.append((label, "row", "present", "missing"))
             continue
         m = re.search(phy + r", /\*addr\*/ (-?\d+), /\*mdc\*/ (-?\d+), /\*mdio\*/ (-?\d+),"
-                      r"\s*/\*rst\*/ (-?\d+), /\*rmiiClk\*/ (-?\d+)", header)
+                      r"\s*/\*rst\*/ (-?\d+), /\*rmiiClk\*/ (-?\d+), /\*extIn\*/ (\w+)", header)
         if not m:
             drift.append((label, phy, "an ethConfigDefault arm", "not found"))
             continue
-        # row is label-less here: [type, phyAddr, mdc, mdio, rst, rmiiClk, ...]
+        # The PHY the preset selects must be the one whose default arm it restates.
+        if row[0] != PHY_VALUE.get(phy, row[0]):
+            drift.append((label, "phyType", PHY_VALUE[phy], row[0]))
+        # row is label-less here: [type, phyAddr, mdc, mdio, rst, rmiiClk, extIn, ...]
         for i, field in enumerate(FIELDS):
             want, got = m.group(i + 1), row[i + 1]
             if want != got:
@@ -310,8 +316,10 @@ def _gpio(v):
             # are genuine BOARD WIRING are required — MDC/MDIO may stay at the IDF default (omit or
             # -1) on RMII, since that's a real standard, not a board-specific value.
             if mtype == "NetworkModule" and isinstance(controls, dict):
-                board = controls.get("ethBoard")
-                if board is not None:
+                # PRESENCE, not truthiness: a JSON `null` reads as None like a missing key, and the
+                # installer treats a present-but-null value as a named preset (null !== "Custom").
+                board = controls["ethBoard"] if "ethBoard" in controls else None
+                if "ethBoard" in controls:
                     if not isinstance(board, str):
                         errors.append(f"{where}: NetworkModule ethBoard must be the preset LABEL as a string, got {board!r}")
                     elif board not in eth_presets:
diff --git a/moondeck/check/check_docgen.py b/moondeck/check/check_docgen.py
index 4486b95b..1a72513f 100755
--- a/moondeck/check/check_docgen.py
+++ b/moondeck/check/check_docgen.py
@@ -1354,6 +1354,51 @@ def _duplicate_group_ids():
     return out
 
 
+# Markdown prose is hard-wrapped when a paragraph line continues onto the next. The standards
+# forbid it for the same reason the header rule does: the editor soft-wraps, so a one-word edit
+# is a one-word diff rather than a reflowed paragraph. Checked over AUTHORED pages only, since a
+# generated one carries its source's shape and is fixed at the generator.
+_MD_SKIP_PREFIX = ("#", "|", "-", "*", ">", "<", "```", "    ", "\t", "!", "=", ":")
+_MD_NUMBERED = re.compile(r"^\d+[.)]\s")
+
+
+def _md_generated(text: str) -> bool:
+    """A page a tool rewrites, which every generator marks with the same warning.
+
+    Keyed on "do not edit by hand" rather than on "generated by": a hand-written page may well
+    say what produced the files it lists, and exempting on that wording let an authored index
+    out of the check. A tool that owns a file says so in the imperative.
+    """
+    head = "\n".join(text.split("\n")[:8]).lower()
+    return "do not edit by hand" in head
+
+
+def _md_hard_wraps(rel: str, text: str):
+    """Each paragraph line that continues onto the next one."""
+    if _md_generated(text):
+        return []
+    out = []
+    lines = text.split("\n")
+    fenced = False
+    for i, line in enumerate(lines[:-1]):
+        stripped = line.strip()
+        if stripped.startswith("```"):
+            fenced = not fenced
+            continue
+        if fenced or not stripped:
+            continue
+        if stripped.startswith(_MD_SKIP_PREFIX) or _MD_NUMBERED.match(stripped):
+            continue
+        nxt = lines[i + 1].strip()
+        if not nxt or nxt.startswith(_MD_SKIP_PREFIX) or _MD_NUMBERED.match(nxt):
+            continue
+        # A line ending mid-sentence is the wrap; one ending a sentence is a deliberate break.
+        if stripped[-1] not in ".!?:":
+            out.append((f"{rel}::line {i + 1}",
+                        "hard wrap: one line per paragraph, let the editor soft-wrap"))
+    return out
+
+
 def _violations():
     out = []
     for rel in _pages():
@@ -1367,6 +1412,13 @@ def _violations():
         for card in _cards(text):
             out.extend(_card_rules(rel, card))
 
+    for rel in sorted((ROOT / "docs").rglob("*.md")):
+        parts = rel.relative_to(ROOT / "docs").parts
+        # moxygen writes generated pages; docs/work/ is the exemption the standards name.
+        if "moxygen" in parts or parts[0] == "work":
+            continue
+        out.extend(_md_hard_wraps(rel.relative_to(ROOT).as_posix(), rel.read_text()))
+
     for rel in _headers():
         out.extend(_header_rules(str(rel), (ROOT / rel).read_text()))
     out.extend(_duplicate_group_ids())
@@ -1543,6 +1595,20 @@ def _blocking(page, items):
                 lo, hi = min(counts[p] for p in tail), max(counts[p] for p in tail)
                 span = f"{lo}" if lo == hi else f"{lo}-{hi}"
                 out.append(f"| {span} each | *{len(tail)} more {kind} files, {n} findings* | |")
+        # Every ERRORING file's findings in full, right where the work is planned. An error
+        # blocks a commit, so the one question its reader has is what to fix, and a summary
+        # sends them back to a stdout that has scrolled away. Warnings stay counted rather
+        # than listed: 3135 of them is the log this page was rewritten to stop being.
+        for page in sorted(files):
+            errs_f = [f for f in files[page] if _blocks(page, f[1])]
+            if not errs_f:
+                continue
+            out += [f"#### `{page}`", ""]
+            for key, why in sorted(errs_f, key=lambda f: f[1]):
+                where = key.rpartition("::")[2] if "::" in key else ""
+                out.append(f"- {where + ': ' if where else ''}{why}")
+            out += [""]
+
         mix = Counter(_rule_name(why) for k, v in files.items()
                       for _, why in v if _blocks(k, why))
         mix_w = Counter(_rule_name(why) for k, v in files.items()
@@ -1625,24 +1691,24 @@ def _ratchet(found) -> list:
     # and comparing the two key spaces silently compares nothing.
     now = Counter(_rule_label(_rule_name(why)) for key, why in found if not _blocks(key, why))
     now["(total)"] = sum(1 for key, why in found if not _blocks(key, why))
-    return [(rule, was, now.get(rule, 0))
-            for rule, was in base.items() if now.get(rule, 0) > was]
+    # The UNION, not the baseline alone: the report lists only rules the run found, so a rule at
+    # zero is absent from `base` and iterating it let that rule rise silently while the total fell.
+    # A rule the baseline never saw starts at zero, which is what "only shrinks" means for it too.
+    return [(rule, base.get(rule, 0), now.get(rule, 0))
+            for rule in sorted(set(base) | set(now)) if now.get(rule, 0) > base.get(rule, 0)]
 
 
 def main() -> int:
     ap = argparse.ArgumentParser(description=__doc__)
-    ap.add_argument("--noreport", action="store_true",
-                    help="skip writing docs/reference/metrics/docgen.md (it is written by default, "
-                         "because the tracked state of the sweep is what gets reviewed)")
-    args = ap.parse_args()
+    ap.parse_args()
 
     found = _violations()
 
     # The report is the artifact: stdout scrolls away and is truncated, the file is what the
-    # reviewer reads. Written on every run unless explicitly suppressed, and the run still
-    # prints and still returns its exit code, so writing it costs a caller nothing.
-    if not args.noreport:
-        _write_report(found)
+    # reviewer reads, and its git history is the only record of the sweep's trend. Written on
+    # every run with no way to suppress it: the run still prints and still returns its exit
+    # code, so writing it costs a caller nothing and a skipped write loses the number to beat.
+    _write_report(found)
 
     errors = [f for f in found if _blocks(f[0], f[1])]
     warnings = [f for f in found if not _blocks(f[0], f[1])]
diff --git a/moondeck/docs/gen_api.py b/moondeck/docs/gen_api.py
index bafb0728..2701ee21 100644
--- a/moondeck/docs/gen_api.py
+++ b/moondeck/docs/gen_api.py
@@ -128,6 +128,12 @@ def _doxyfile(headers: list[str], xml_out: str) -> str:
         "JAVADOC_AUTOBRIEF = YES\n"          # a leading `///`/`//` line is the brief
         f"STRIP_FROM_PATH = {ROOT}\n"        # relative "Defined in src/…", never an abs path
         "QUIET = YES\nWARN_IF_UNDOCUMENTED = NO\n"
+        # A capability macro the compiler sets per target, which Doxygen never sees: left
+        # undefined it preprocesses the guarded code away, and a header that is entirely
+        # inside one documents nothing at all. Defined here so the page exists on every
+        # build, the same way the widest target compiles it.
+        "ENABLE_PREPROCESSING = YES\nMACRO_EXPANSION = YES\n"
+        "PREDEFINED = MM_HEAVY_COMPUTE=1\n"
     )
 
 
@@ -347,16 +353,30 @@ def _relocate_moreinfo(md: str) -> str:
 
 def _render_card_directives(md: str, domain: str, stem: str) -> str:
     """Replace each `@card <file>` directive with an <img> pointing at the resolved asset.
-    A file that doesn't exist on disk drops the directive (no broken image) — the same
-    fail-soft as a missing generated page."""
+
+    A moving preview names a `.gif`, and the module's own card screenshot sits beside it under the
+    same stem: both are emitted, the animation first and the controls that produced it second, so a
+    reader sees what it looks like and what it is made of. A file absent from disk drops silently,
+    the same fail-soft as a missing generated page."""
     def repl(m: re.Match) -> str:
         fname = m.group("file")
         # Search under docs/assets/<domain>/ for the file (handles the light/{drivers,effects,…} subdir).
         hits = list((_ASSETS / domain).rglob(fname))
         if not hits:
             return ""   # asset absent → emit nothing rather than a broken link
-        rel = os.path.relpath(hits[0], DOCS_MOONMODULES / domain / "moxygen")
-        return f'\n\n<img src="{rel}" alt="{stem} card" width="300">\n'
+        shots = [hits[0]]
+        # The controls screenshot pairs with a moving preview, and only there: a card that already
+        # IS the screenshot has no second image to add.
+        if hits[0].suffix == ".gif":
+            shot = hits[0].with_suffix(".png")
+            if shot.exists():
+                shots.append(shot)
+        out = ""
+        for i, sh in enumerate(shots):
+            rel = os.path.relpath(sh, DOCS_MOONMODULES / domain / "moxygen")
+            alt = f"{stem} card" if i == 0 else f"{stem} controls"
+            out += f'<img src="{rel}" alt="{alt}" width="300">\n'
+        return f"\n\n{out}"
     return _CARD_RE.sub(repl, md)
 
 
@@ -599,10 +619,34 @@ def generate() -> dict[str, str]:
                 else:
                     cls_blocks.setdefault(header, []).append(text)
 
+        # A class DECLARED OUTSIDE the group is absent from the group file, so appending its
+        # per-class block adds it rather than duplicating it. Keyed on the rendered heading: a
+        # class the group already covers is skipped, one it never saw is kept. Without this a
+        # mixed header, free functions in a group plus a class beside it, published the functions
+        # and silently dropped the class, and the card linking it arrived at a page without it.
         by_header: dict[str, list[str]] = {h: [md] for h, md in grp_headers.items()}
         for header, blocks in cls_blocks.items():
             if header not in by_header:
                 by_header[header] = blocks
+                continue
+            grp = by_header[header][0]
+            # The group file's OWN lead is its title, which for a `@defgroup JsonSink` is the class
+            # name itself: matching against it would read every such class as already rendered.
+            # So compare against the headings BELOW that lead, which are what the group really
+            # contains, and keep a class the group never rendered.
+            # The group's own `# Title` line is dropped before the comparison: for a
+            # `@defgroup JsonSink` it carries the class's name and would read as the class
+            # already being rendered. What follows it is what the group really contains.
+            lines = grp.split("\n")
+            for i, ln in enumerate(lines):
+                if ln.startswith("# "):
+                    lines = lines[i + 1:]
+                    break
+            grp_members = set(re.findall(r'^#+ (\S+)', "\n".join(lines), re.M))
+            for b in blocks:
+                m = re.search(r'^#+ (\S+)', b, re.M)
+                if m and m.group(1) not in grp_members:
+                    by_header[header].append(b)
 
         pages: dict[str, str] = {}
         for header, blocks in by_header.items():
diff --git a/moondeck/docs/generate_test_docs.py b/moondeck/docs/generate_test_docs.py
index e560523f..545104f9 100644
--- a/moondeck/docs/generate_test_docs.py
+++ b/moondeck/docs/generate_test_docs.py
@@ -8,7 +8,7 @@
 `also`, `name`, `description`, and per-step `description`.
 
 Both outputs are grouped by primary `@module` / `module`. The script is the
-single owner of the generated files — running it idempotently produces the
+single owner of the generated files, running it idempotently produces the
 same bytes (verified by --check).
 
 Usage:
@@ -46,7 +46,7 @@ def render_unit_tests(files: list[dict]) -> str:
     lines.append("")
     lines.append(
         "Auto-generated from `test/unit/{core,light}/unit_*.cpp` by `moondeck/docs/generate_test_docs.py`. "
-        "**Do not edit by hand** — update the source file's `@module` / `@also` "
+        "**Do not edit by hand**: update the source file's `@module` / `@also` "
         "and per-TEST_CASE `//` descriptions instead, then regenerate."
     )
     lines.append("")
@@ -63,7 +63,7 @@ def render_unit_tests(files: list[dict]) -> str:
         for f in by_module[module]:
             rel = f["path"].relative_to(ROOT).as_posix()
             if f["file_description"]:
-                lines.append(f"`{rel}` — {f['file_description']}")
+                lines.append(f"`{rel}`: {f['file_description']}")
             else:
                 lines.append(f"`{rel}`")
             if f["also"]:
@@ -213,7 +213,7 @@ def _format_perf_table(step: dict) -> list[str]:
         return []
 
     lines: list[str] = []
-    lines.append("**Performance** (contract / observed) — tick stored, FPS shown:")
+    lines.append("**Performance** (contract / observed), tick stored, FPS shown:")
     lines.append("")
     lines.append("| Board | FPS | heap | block |")
     lines.append("|---|---|---|---|")
@@ -250,7 +250,7 @@ def _format_perf_table(step: dict) -> list[str]:
 def _fmt_us_range(v) -> str:
     """Pretty-print a [min, max] tick range. Collapses when the *formatted*
     endpoints would render identically (e.g. 84,500µs and 84,520µs both round
-    to "85µs" at our resolution — showing them as a range adds noise)."""
+    to "85µs" at our resolution, showing them as a range adds noise)."""
     if isinstance(v, list) and len(v) == 2:
         lo, hi = int(v[0]), int(v[1])
         lo_s, hi_s = _fmt_us(lo), _fmt_us(hi)
@@ -311,7 +311,7 @@ def render_scenarios(files: list[dict]) -> str:
     lines.append("")
     lines.append(
         "Auto-generated from `test/scenarios/{core,light}/scenario_*.json` by "
-        "`moondeck/docs/generate_test_docs.py`. **Do not edit by hand** — "
+        "`moondeck/docs/generate_test_docs.py`. **Do not edit by hand**: "
         "update the JSON file's top-level fields and per-step `description` "
         "/ `bounds` / `contract` / `observed` instead, then regenerate."
     )
@@ -334,7 +334,7 @@ def render_scenarios(files: list[dict]) -> str:
             rel = f["path"].relative_to(ROOT).as_posix()
             lines.append(f"### {f['name']}")
             lines.append("")
-            lines.append(f"`{rel}` — {f['description']}")
+            lines.append(f"`{rel}`: {f['description']}")
             lines.append("")
             # Top-level scenario flags worth surfacing.
             meta_bits: list[str] = [f"**Mode**: `{f['mode']}`"]
@@ -368,7 +368,7 @@ def render_scenarios(files: list[dict]) -> str:
                         bits = [f"`{p['name']}` ({p['op']})"]
                         if p["description"]:
                             bits.append(p["description"])
-                        lines.append(f"- {' — '.join(bits)}")
+                        lines.append(f"- {': '.join(bits)}")
                     lines.append("")
                     prep_buffer = []
                 bounds = _format_bounds(step["bounds"])
@@ -387,7 +387,7 @@ def render_scenarios(files: list[dict]) -> str:
                     bits = [f"`{p['name']}` ({p['op']})"]
                     if p["description"]:
                         bits.append(p["description"])
-                    lines.append(f"- {' — '.join(bits)}")
+                    lines.append(f"- {': '.join(bits)}")
                 lines.append("")
 
     return "\n".join(lines).rstrip() + "\n"
diff --git a/moondeck/docs/screenshot_modules.py b/moondeck/docs/screenshot_modules.py
index 885edf7a..a6547cd4 100644
--- a/moondeck/docs/screenshot_modules.py
+++ b/moondeck/docs/screenshot_modules.py
@@ -135,6 +135,10 @@ def asset_dir_for(type_name: str) -> Path:
     # silicon or the runtime is absent: HUB75 wants LCD_CAM or Parlio, NDI wants a
     # runtime the user installs. The card and its controls are what a screenshot is
     # for, and those are real on any host.
+    # The two LED drivers, each its own card: they share only `pins` and `ledsPerPin`, and a
+    # screenshot of one describes the other's controls wrongly.
+    ("RmtLedDriver",        "Drivers",  {}, False),
+    ("ParallelLedDriver",   "Drivers",  {}, False),
     ("Hub75Driver",         "Drivers",  {}, False),
     ("NdiDriver",           "Drivers",  {}, False),
     ("HlsDriver",           "Drivers",  {}, False),
diff --git a/src/core/system/MoonCloudModule.h b/src/core/system/MoonCloudModule.h
index 67254708..2a47657e 100644
--- a/src/core/system/MoonCloudModule.h
+++ b/src/core/system/MoonCloudModule.h
@@ -2,13 +2,6 @@
 #pragma once
 
 
-/// @defgroup MoonCloud The container for everything projectMM does with a server we run
-/// @{
-///
-/// It holds no controls and does no work of its own.
-/// Each thing MoonCloud does is a child with its own consent: Stats, Talk, and Sync to come.
-/// A user who wants a joint lightshow has not agreed to usage reporting.
-
 #include <cstddef>
 #include <cstdint>
 #include <cstdio>
@@ -20,6 +13,11 @@
 
 namespace mm {
 
+/// The container for everything projectMM does with a server we run.
+///
+/// It holds no controls and does no work of its own.
+/// Each thing MoonCloud does is a child with its own consent: Stats, Talk, and Sync to come.
+/// A user who wants a joint lightshow has not agreed to usage reporting.
 class MoonCloudModule : public MoonModule {
 public:
     /// Always runs: the children carry the real consent, so disabling the container is ambiguous.
@@ -43,7 +41,10 @@ class MoonCloudModule : public MoonModule {
 };
 
 
-// A salted hash of the MAC: pseudonymous, and salted apart from the MQTT and Home Assistant ids.
+/// @defgroup MoonCloudIdentity The installation id, a salted hash of the MAC
+/// @{
+///
+/// Pseudonymous, and salted apart from the MQTT and Home Assistant ids.
 
 /// Characters written by `installationId`, excluding the terminator.
 inline constexpr size_t kInstallationIdChars = 32;
diff --git a/src/core/system/NetworkModule.h b/src/core/system/NetworkModule.h
index 7d91b496..56370d3c 100644
--- a/src/core/system/NetworkModule.h
+++ b/src/core/system/NetworkModule.h
@@ -552,7 +552,14 @@ class NetworkModule : public MoonModule {
     static bool presetBuildable(const EthPreset& p) {
         if (p.type == 0) return true;                       // Custom, which names no PHY
         if (p.type == 3) return platform::hasEthW5500;      // SPI, a separate driver
-        return !platform::hasEthW5500;                      // the internal EMAC drives the rest
+        if (!platform::hasEthW5500) {
+            // A preset carries a PIN MAP as well as a PHY, so one the chip cannot wire persists pins that reach nothing. The desktop previews the catalog and sees them all.
+            constexpr bool knownChip = platform::isEsp32P4 || platform::isEsp32S31;
+            if (p.type == 2) return !knownChip || platform::isEsp32P4;    // IP101: the P4's
+            if (p.type == 4) return !knownChip || platform::isEsp32S31;   // YT8531 RGMII: the S31's
+            return !platform::isEsp32P4 && !platform::isEsp32S31;         // LAN8720: classic RMII
+        }
+        return false;
     }
 
     /// Offer the presets this build can drive, re-pointing the selection by label.
diff --git a/src/light/drivers/PreviewDriver.h b/src/light/drivers/PreviewDriver.h
index 532ffaf6..a57afa11 100644
--- a/src/light/drivers/PreviewDriver.h
+++ b/src/light/drivers/PreviewDriver.h
@@ -72,6 +72,32 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
     }
 
 
+
+    /// The stride a color frame ships at: the link's pacing, never finer than the table's.
+    nrOfLightsType frameStride() const {
+        return downscale_ > previewStride_ ? downscale_ : previewStride_;
+    }
+
+    /// How many lights a frame at `frameStride()` carries, counted the way the table is.
+    nrOfLightsType frameCount() const {
+        const nrOfLightsType fs = frameStride();
+        if (fs == previewStride_) return coordCount_;
+        if (!layer_) return coordCount_;
+        const lengthType ax = layer_->physicalWidth()  > 0 ? layer_->physicalWidth()  : 1;
+        const lengthType ay = layer_->physicalHeight() > 0 ? layer_->physicalHeight() : 1;
+        const lengthType az = layer_->physicalDepth()  > 0 ? layer_->physicalDepth()  : 1;
+        if (denseGrid()) {
+            const nrOfLightsType cx = (ax + fs - 1) / fs, cy = (ay + fs - 1) / fs,
+                                 cz = (az + fs - 1) / fs;
+            return static_cast<nrOfLightsType>(static_cast<uint32_t>(cx) * cy * cz);
+        }
+        // From the CACHE, never a placeLights walk: this runs per frame on a nonblocking tick.
+        if (!keptStep_ || keptCount_ != coordCount_) return coordCount_;
+        nrOfLightsType out = 0;
+        for (nrOfLightsType k = 0; k < keptCount_; k++) if (keptStep_[k] % fs == 0) out++;
+        return out;
+    }
+
     /// Test-only: the currently served downsample factor, 1 being full resolution.
     nrOfLightsType downscaleForTest() const { return downscale_; }
 
@@ -106,11 +132,17 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
             ensureStaging(maxPts * 3u);
             if (!denseGrid() && keptIdxCap_ < maxPts) {
                 auto* grown = static_cast<nrOfLightsType*>(platform::alloc(maxPts * sizeof(nrOfLightsType)));
-                if (grown) {
+                auto* steps = static_cast<uint8_t*>(platform::alloc(maxPts));
+                if (grown && steps) {
                     if (keptIdx_) platform::free(keptIdx_);
+                    if (keptStep_) platform::free(keptStep_);
                     keptIdx_ = grown;
+                    keptStep_ = steps;
                     keptIdxCap_ = static_cast<nrOfLightsType>(maxPts);
                     publishHeapBytes();
+                } else {
+                    if (grown) platform::free(grown);
+                    if (steps) platform::free(steps);
                 }
             }
         }
@@ -222,7 +254,7 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
             };
             while (latticeCount(s) > cap) s++;
         }
-        if (s < downscale_) s = downscale_;   // adaptive: never finer than the link sustains
+        // MEMORY bounds the table, never the link: geometry is sent once, and `downscale_` paces the colors instead.
         previewStride_ = s;
 
         // A dense grid is closed-form; a mapped layout is counted by one placeLights pass.
@@ -241,13 +273,18 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
             // Sized to EXACTLY this count before the emit fills it, so the cache cannot truncate.
             if (keptIdxCap_ < coordCount_) {
                 auto* grown = static_cast<nrOfLightsType*>(platform::alloc(coordCount_ * sizeof(nrOfLightsType)));
-                if (grown) {
+                auto* steps = static_cast<uint8_t*>(platform::alloc(coordCount_));
+                if (grown && steps) {
                     if (keptIdx_) platform::free(keptIdx_);
+                    if (keptStep_) platform::free(keptStep_);
                     keptIdx_ = grown;
+                    keptStep_ = steps;
                     keptIdxCap_ = coordCount_;
                     keptIdxAllocFailed_ = false;
                     publishHeapBytes();   // the index cache grew: refresh the memory readout
                 } else {
+                    if (grown) platform::free(grown);
+                    if (steps) platform::free(steps);
                     keptIdxAllocFailed_ = true;   // degraded: the gather walks placeLights per frame
                 }
             }
@@ -280,8 +317,14 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
                 auto* p = static_cast<PosCtx*>(c);
                 if (x % p->s != 0 || y % p->s != 0 || z % p->s != 0) return;
                 PreviewDriver* self = p->self;
-                if (self->keptIdx_ && self->keptCount_ < self->keptIdxCap_)
+                if (self->keptIdx_ && self->keptCount_ < self->keptIdxCap_) {
+                    // The coarsest lattice this light sits on: every axis divisible by the step.
+                    uint8_t step = 1;
+                    for (uint8_t t = 2; t <= 64; t++)
+                        if (x % t == 0 && y % t == 0 && z % t == 0) step = t;
+                    if (self->keptStep_) self->keptStep_[self->keptCount_] = step;
                     self->keptIdx_[self->keptCount_++] = idx;
+                }
                 p->emit(x, y, z);
             }, nullptr, &pc});
         }
@@ -373,30 +416,32 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
         const uint8_t* src = sourceBuffer_->data();
         const uint8_t cpl = sourceBuffer_->channelsPerLight();
         const nrOfLightsType n = sourceBuffer_->count();
-        const nrOfLightsType s = previewStride_;
+        // The LINK's stride, coarser than the table's when the link is slow: the browser upsamples the colors back onto the full table it already holds.
+        const nrOfLightsType s = frameStride();
+        const nrOfLightsType sendCount = frameCount();
 
         // The epoch and stride pair is the client's table-cache key; drops is its congestion signal.
         uint8_t header[9];
         header[0] = 0x02;
-        header[1] = static_cast<uint8_t>(coordCount_ & 0xFF);
-        header[2] = static_cast<uint8_t>((coordCount_ >> 8) & 0xFF);
-        header[3] = static_cast<uint8_t>((coordCount_ >> 16) & 0xFF);
-        header[4] = static_cast<uint8_t>((coordCount_ >> 24) & 0xFF);
+        header[1] = static_cast<uint8_t>(sendCount & 0xFF);
+        header[2] = static_cast<uint8_t>((sendCount >> 8) & 0xFF);
+        header[3] = static_cast<uint8_t>((sendCount >> 16) & 0xFF);
+        header[4] = static_cast<uint8_t>((sendCount >> 24) & 0xFF);
         header[5] = static_cast<uint8_t>(s & 0xFF);
         header[6] = static_cast<uint8_t>(s >> 8);
         header[7] = epoch_;
         header[8] = dropsSinceLast_;
 
-        if (s == 1 && cpl == 3 && coordCount_ <= n) {
+        if (s == 1 && cpl == 3 && sendCount <= n) {
             // Full resolution: the producer buffer IS the payload, drained with no copy at all.
             const bool ok = broadcaster_->sendBufferedFrame(header, sizeof(header),
-                                                            src, static_cast<size_t>(coordCount_) * 3);
+                                                            src, static_cast<size_t>(sendCount) * 3);
             if (ok) dropsSinceLast_ = 0;
             return ok;
         }
 
         // Gathered into staging, in the coord table's exact subset and order, or the browser drops it.
-        const size_t bodyBytes = static_cast<size_t>(coordCount_) * 3;
+        const size_t bodyBytes = static_cast<size_t>(sendCount) * 3;
         if (!staging_ || stagingCap_ < bodyBytes) return false;   // alloc miss: skip, lossy channel
         struct ColCtx {
             uint8_t* out; size_t at; const uint8_t* src; nrOfLightsType n; uint8_t cpl;
@@ -418,8 +463,9 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
                         col.emit(static_cast<nrOfLightsType>(static_cast<size_t>(z) * H * W
                                                              + static_cast<size_t>(y) * W + x));
         } else if (keptIdx_ && keptCount_ == coordCount_) {
-            // The index map cached at coord-table build: a tight gather over the kept lights only.
-            for (nrOfLightsType k = 0; k < keptCount_; k++) col.emit(keptIdx_[k]);
+            // The cached index map, filtered by the frame's stride: a light survives when that stride divides the coarsest lattice it sits on.
+            for (nrOfLightsType k = 0; k < keptCount_; k++)
+                if (!keptStep_ || s == previewStride_ || keptStep_[k] % s == 0) col.emit(keptIdx_[k]);
         } else {
             // The alloc-miss fallback: the full lattice walk, at the same stride as the table's.
             struct Skip { ColCtx* col; nrOfLightsType s; } sk{&col, s};
@@ -439,6 +485,7 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
     void freePreviewBuffers() {
         if (broadcaster_) broadcaster_->cancelBufferedSend();
         if (keptIdx_) { platform::free(keptIdx_); keptIdx_ = nullptr; keptIdxCap_ = 0; keptCount_ = 0; }
+        if (keptStep_) { platform::free(keptStep_); keptStep_ = nullptr; }
         if (staging_) { platform::free(staging_); staging_ = nullptr; stagingCap_ = 0; }
         publishHeapBytes();
     }
@@ -487,6 +534,8 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
     bool keptIdxAllocFailed_ = false;     // index cache couldn't allocate → gather walks per frame
     nrOfLightsType* keptIdx_ = nullptr;   // sparse layouts: kept lights' buffer indices, coord-table order
     nrOfLightsType keptIdxCap_ = 0, keptCount_ = 0;
+    // The coarsest stride each kept light survives, so a coarser frame filters the cache rather than re-walking placeLights (~8 ms at 12K lights, per frame).
+    uint8_t* keptStep_ = nullptr;
 
 protected:
     /// This driver's heap: the base scratch plus the kept-index cache, for the memory readout.
diff --git a/src/light/effects/EchoEffect.h b/src/light/effects/EchoEffect.h
index b47d5532..f69ee5fc 100644
--- a/src/light/effects/EchoEffect.h
+++ b/src/light/effects/EchoEffect.h
@@ -36,7 +36,7 @@ class EchoEffect : public EffectBase {
     uint8_t bpm    = 30;
     /// How much the feedback grows each pass, where 0 leaves it still.
     uint8_t zoom   = 12;
-    /// Rotation per pass, which turns the trail into a spiral.
+    /// Rotation per frame, which turns the trail into a spiral.
     uint8_t rotate = 8;
     /// How fast the echo fades, so higher gives a shorter trail.
     uint8_t decay  = 24;
diff --git a/src/ui/app.js b/src/ui/app.js
index c78e5d6c..d49f828a 100644
--- a/src/ui/app.js
+++ b/src/ui/app.js
@@ -993,6 +993,7 @@ function uiMode() {
 function renderCards() {
     const main = document.getElementById("main");
     if (!main || !state) return;
+    renderedMode_ = uiMode();   // what these cards are being built for
     main.innerHTML = "";
 
     // One root visible at a time: render only the selected root's subtree.
@@ -1838,6 +1839,9 @@ function createCard(mod, depth) {
         stats.addEventListener("mouseleave", hide);
         stats.addEventListener("touchstart", (e) => { e.preventDefault(); show(); });
         stats.addEventListener("touchend", hide);
+        // A cancelled touch (a scroll takes over, a call arrives) fires no touchend, which left
+        // the peek latched until the next gesture.
+        stats.addEventListener("touchcancel", hide);
         title.appendChild(stats);
     }
 
@@ -2563,8 +2567,15 @@ function createControl(moduleName, moduleType, ctrl) {
     // get a distinct treatment so they read as a different tier: a left accent stripe + muted label.
     // One class per tier, because the glyph says WHICH mode reveals it: a reader in expert mode
     // meets both marks and the developer one tells them what they are still not seeing.
-    if ((ctrl.minMode | 0) === 1) row.classList.add("control-expert");
-    else if ((ctrl.minMode | 0) >= 2) row.classList.add("control-developer");
+    // The glyph is CSS ::after content, which a screen reader does not reliably announce, so the
+    // tier also rides a title: the same fact reaches a reader who cannot see the mark.
+    if ((ctrl.minMode | 0) === 1) {
+        row.classList.add("control-expert");
+        row.title = "Expert mode control";
+    } else if ((ctrl.minMode | 0) >= 2) {
+        row.classList.add("control-developer");
+        row.title = "Developer mode control";
+    }
     // A switch-row control renders as a strip like an encoder or fader, so switch N sits in the
     // same column as encoder N and fader N: a surface reads down a channel, not across a list.
     if (ctrl.switchRow) row.classList.add("control-switch");
@@ -2956,6 +2967,7 @@ function createControl(moduleName, moduleType, ctrl) {
             peek.addEventListener("mouseleave", hide);
             peek.addEventListener("touchstart", (e) => { e.preventDefault(); show(); });
             peek.addEventListener("touchend", hide);
+            peek.addEventListener("touchcancel", hide);   // a cancelled touch fires no touchend
             row.appendChild(peek);
             break;
         }
@@ -4461,8 +4473,21 @@ function setText(el, text) {
     if (el && el.textContent !== text) el.textContent = text;
 }
 
+// The mode the cards were BUILT at. `uiMode()` decides which controls and which stats line a card
+// gets, so a change to it reshapes the DOM rather than a value in it: the one case the no-rebuild
+// contract cannot patch. Tracked here because the change can come from another client entirely.
+let renderedMode_ = null;
+
 function updateValues() {
     if (!state || !state.modules) return;
+    // A mode change adds or removes rows, so patching values cannot express it: rebuild once.
+    const mode = uiMode();
+    if (renderedMode_ !== null && renderedMode_ !== mode) {
+        renderedMode_ = mode;
+        renderCards();
+        return;
+    }
+    renderedMode_ = mode;
     // Patch each visible card's controls and stats line; never rebuild the DOM here.
     for (const mod of allModules()) {
         updateTabDot(mod);   // a fault on a BACKGROUND tab must surface without opening it
diff --git a/src/ui/preview3d.js b/src/ui/preview3d.js
index 1ebb5d3e..d897d315 100644
--- a/src/ui/preview3d.js
+++ b/src/ui/preview3d.js
@@ -125,6 +125,9 @@ const clampDot = (v) => Math.min(DOT_MAX, Math.max(DOT_MIN, Number.isFinite(v) ?
 let dotScale_ = clampDot(parseFloat(localStorage.getItem("mm_preview_dot")));
 let resetLayout_ = null;     // set by setupLayout(): restores docked/PiP state to defaults
 let previewBox_ = null;      // {x,y,z} bounding-box extent for camera auto-fit
+let previewPos_ = null;      // the table's raw (x,y,z) bytes: what the stride test reads
+let previewTableStride_ = 1; // the stride the CACHED table was built at, the finest available
+let previewEpochKey_ = 0;    // the epoch the cached table belongs to, keying the upsample map
 let lineProgram = null;      // separate program for the wireframe bounding box
 let lineLocs = null;
 let lineBuffer = null;
@@ -631,11 +634,12 @@ function parsePreviewCoords(view, buf) {
         coords[i * 3 + 1] = (pos[i * 3 + 1] / maxDim) - 0.5 * by / maxDim;
         coords[i * 3 + 2] = (pos[i * 3 + 2] / maxDim) - 0.5 * bz / maxDim;
     }
-    // CACHE the table per (epoch, stride): a stride change back to a cached rung costs zero table
-    // traffic, the lean channel's core idea. Tables from dead epochs are dropped (the device
-    // renumbered the world); browser memory for one epoch's whole ladder is a few hundred KB.
+    // ONE table per epoch, the finest the device will serve. Every coarser stride is derived from
+    // it here rather than fetched, so a stride change costs no table traffic at any rung and the
+    // preview keeps the layout's full extent at every detail level. Positions are static for an
+    // epoch, so this is also less traffic than a table per rung: one fetch, not three or four.
     for (const k of tableCache_.keys()) if (!k.startsWith(epoch + ":")) tableCache_.delete(k);
-    tableCache_.set(epoch + ":" + stride, { coords, count, maxDim, bx, by, bz });
+    tableCache_.set(epoch + ":" + stride, { coords, count, maxDim, bx, by, bz, stride, pos: new Uint8Array(pos) });
     activateTable(epoch, stride);
     // Draw the geometry NOW, dark, so a fresh page shows the layout the instant the table arrives,
     // not only once the first color frame lands. Color frames then light it.
@@ -645,17 +649,60 @@ function parsePreviewCoords(view, buf) {
 // Make a cached table the rendering one. Returns false when the cache misses (the caller then
 // asks the device for it: the pull model).
 function activateTable(epoch, stride) {
-    const t = tableCache_.get(epoch + ":" + stride);
+    // Any cached table for this epoch serves every stride: the colors are mapped onto it below.
+    const t = tableCache_.get(epoch + ":" + stride)
+           || [...tableCache_.keys()].filter(k => k.startsWith(epoch + ":"))
+                .map(k => tableCache_.get(k))[0];
     if (!t) return false;
     previewCoords_ = t.coords;
     previewCoordCount_ = t.count;
     previewMaxDim_ = t.maxDim;
     previewBox_ = { x: t.bx, y: t.by, z: t.bz };
+    previewPos_ = t.pos;
+    previewTableStride_ = t.stride ?? 1;
+    previewEpochKey_ = epoch;
     previewStride_ = stride;
     updatePreviewStatus();
     return true;
 }
 
+// Which table point each color in a stride-`s` frame belongs to. The device keeps a light when
+// `x % s == 0 && y % s == 0 && z % s == 0` (PreviewDriver::buildCoordTable), walking z, then y,
+// then x, so running the same test over the table's own positions reproduces its order exactly.
+// Every point then takes the color of the kept light leading its block: the layout stays whole
+// and only the detail coarsens. Cached per (table, stride), since it changes only when one does.
+let upsampleCache_ = new Map();
+function upsampleMap(stride) {
+    const base = previewTableStride_ || 1;
+    const n = previewCoordCount_;
+    if (!previewPos_ || stride <= base) return null;   // already the finest the device serves
+    const key = previewEpochKey_ + ":" + base + ":" + stride;
+    const hit = upsampleCache_.get(key);
+    if (hit && hit.length === n) return hit;
+    // The kept lights, in the device's emit order.
+    const kept = [];
+    for (let i = 0; i < n; i++) {
+        const x = previewPos_[i * 3], y = previewPos_[i * 3 + 1], z = previewPos_[i * 3 + 2];
+        if (x % stride === 0 && y % stride === 0 && z % stride === 0) kept.push(i);
+    }
+    // Each point to the nearest kept light at or before it on each axis: its block's leader.
+    const map = new Uint32Array(n);
+    const slot = new Map();
+    for (let k = 0; k < kept.length; k++) {
+        const i = kept[k];
+        slot.set(previewPos_[i * 3] + "," + previewPos_[i * 3 + 1] + "," + previewPos_[i * 3 + 2], k);
+    }
+    for (let i = 0; i < n; i++) {
+        const bxq = previewPos_[i * 3] - (previewPos_[i * 3] % stride);
+        const byq = previewPos_[i * 3 + 1] - (previewPos_[i * 3 + 1] % stride);
+        const bzq = previewPos_[i * 3 + 2] - (previewPos_[i * 3 + 2] % stride);
+        const k = slot.get(bxq + "," + byq + "," + bzq);
+        map[i] = k === undefined ? 0 : k;
+    }
+    upsampleCache_ = new Map([[key, map]]);   // one epoch, one stride in flight: keep the last
+    return map;
+}
+
 // Ask the device for the coordinate table ([0x52][stride]), at most once per half second: the
 // answer is paced by the drain, and re-asking faster only queues duplicate work.
 function requestTable(stride) {
@@ -682,16 +729,31 @@ function renderPreviewFrame(view, buf) {
     // table traffic); a miss asks the device for the positions and skips this frame, the pull
     // model's whole geometry story.
     if ((epoch !== lastEpoch_ || stride !== previewStride_) && !activateTable(epoch, stride)) {
-        requestTable(stride);
+        // The FINEST table, once per epoch: every coarser stride is derived from it here.
+        requestTable(1);
         return;
     }
-    if (count !== previewCoordCount_) return;   // mid-rebuild mismatch: the next table realigns
+    previewStride_ = stride;
+    const up = upsampleMap(stride);
+    // A frame carries one color per kept light; without a map the two counts must agree.
+    if (!up && count !== previewCoordCount_) return;   // mid-rebuild mismatch: the next table realigns
     const rgb = new Uint8Array(buf, 9);
     // Kept for drawBeams: a moving head's beam is the color the fixture is EMITTING, so the beam
     // pass needs the same frame the dots were drawn from. A COPY, because this outlives the
     // message: drawBeams reads it on every orbit redraw, and a view onto a recycled receive
     // buffer would color beams from whatever arrived next.
-    previewRgb_ = new Uint8Array(rgb);
+    // UPSAMPLE: one color per kept light becomes one per table point, each taking its block's.
+    // The layout keeps its full extent and every light stays on screen; only the detail coarsens.
+    if (up) {
+        const full = new Uint8Array(previewCoordCount_ * 3);
+        for (let i = 0; i < previewCoordCount_; i++) {
+            const src = up[i] * 3;
+            full[i * 3] = rgb[src]; full[i * 3 + 1] = rgb[src + 1]; full[i * 3 + 2] = rgb[src + 2];
+        }
+        previewRgb_ = full;
+    } else {
+        previewRgb_ = new Uint8Array(rgb);
+    }
     // The first aim frame changes what the scene CONTAINS: beams extend well past the fixtures,
     // and a fit measured before they existed frames only the heads. Re-arm the auto-fit once so
     // the next one accounts for them; `sawAim_` keeps it to once, not once per aim frame.
@@ -702,7 +764,7 @@ function renderPreviewFrame(view, buf) {
     // showed beams on every light. The device alternates aim and color frames, so a couple of
     // color frames with no aim between them means the aim stream has ended.
     if (previewAim_ && ++framesSinceAim_ > kAimStaleFrames) { previewAim_ = null; sawAim_ = false; }
-    drawLights(rgb);
+    drawLights(previewRgb_);
     measureFrameRate();
 }
 
diff --git a/test/js/ui-live-patch-text.test.mjs b/test/js/ui-live-patch-text.test.mjs
index dd02a73f..d730cb3b 100644
--- a/test/js/ui-live-patch-text.test.mjs
+++ b/test/js/ui-live-patch-text.test.mjs
@@ -114,8 +114,12 @@ for (const fn of ["updateValues", "updateModuleControls", "updateStatusBar", "ap
 // nobody adds here is exactly how this bug came back twice. Pin that every function updateValues
 // calls into (our own, not DOM builtins) is one this file checks.
 test("every function on the patch path is covered by this file", () => {
+    // renderCards is the ESCAPE from the patch path, not a step on it: a mode change adds or
+    // removes control rows, which patching values cannot express, so the card is rebuilt and
+    // updateValues returns. It collapses a selection by design, which is why it runs only on a
+    // real change rather than every second.
     const checked = ["updateValues", "updateModuleControls", "updateStatusBar", "applyTabDot",
-                     "updateTabDot", "setText", "setUrlDisplay", "setStatusText"];
+                     "updateTabDot", "setText", "setUrlDisplay", "setStatusText", "renderCards"];
     const keywords = ["if", "for", "while", "switch", "catch", "return", "typeof"];
     const dom = ["querySelector", "querySelectorAll", "queryByName", "createElement", "appendChild",
                  "insertBefore", "toggle", "setAttribute", "getAttribute", "remove", "closest",
@@ -124,7 +128,8 @@ test("every function on the patch path is covered by this file", () => {
     const called = [...new Set(Array.from(body.matchAll(/\b([a-zA-Z_][\w]*)\s*\(/g), m => m[1]))]
         .filter(n => !dom.includes(n) && !keywords.includes(n))
         // Formatters return strings and touch no DOM, so they cannot break a selection.
-        .filter(n => !/^(fmt|format|css|now|allModules|previewTargetFps)/.test(n));
+        // uiMode reads one control's value and returns a number: no DOM, like the formatters.
+        .filter(n => !/^(fmt|format|css|now|allModules|previewTargetFps|uiMode)/.test(n));
     const missing = called.filter(n => !checked.includes(n));
     assert.deepEqual(missing, [],
         `these run on every state patch but are not audited here: add them to the lists above ` +
diff --git a/test/python/test_check_docgen.py b/test/python/test_check_docgen.py
index 2dc66945..84765c13 100644
--- a/test/python/test_check_docgen.py
+++ b/test/python/test_check_docgen.py
@@ -951,6 +951,26 @@ def test_the_ratchet_refuses_a_rule_that_rose_against_the_committed_report(monke
     assert counts is None or not any(k.startswith("`") for k in counts), counts
 
 
+def test_the_ratchet_catches_a_rule_the_baseline_never_had(monkeypatch):
+    """A rule at zero is ABSENT from the report, so the baseline has no row for it. Iterating the
+    baseline alone let such a rule rise silently while the total fell, which is the same hole the
+    per-rule half exists to close. The comparison is over the union, a missing baseline reading
+    as zero."""
+    import check_docgen
+    base = {"(total)": 10, "multi-line comment blocks": 4}
+    monkeypatch.setattr(check_docgen, "_committed_counts", lambda: base)
+
+    cpp = "src/core/x.cpp"
+    # A warning-side rule with no baseline row. (A hard wrap cannot serve here: it blocks
+    # everywhere by design, so it is an error and never enters the warning column.)
+    fresh = "file opens with // rather than ///"
+    assert check_docgen._rule_label(check_docgen._rule_name(fresh)) not in base
+
+    rose = check_docgen._ratchet([(cpp, fresh)] * 5)
+    assert rose == [("files opening with //", 0, 5)], \
+        f"a rule absent from the baseline must still only shrink: {rose}"
+
+
 def test_a_hard_wrap_blocks_in_an_implementation_file_too():
     """The one rule staged the OTHER way. Every other `.cpp` finding warns while the sweep runs,
     but the tree is at ZERO hard wraps, so there is nothing left to stage and the exemption only
diff --git a/test/scenarios/core/scenario_MoonModule_control_change.json b/test/scenarios/core/scenario_MoonModule_control_change.json
index 27c460be..c206031f 100644
--- a/test/scenarios/core/scenario_MoonModule_control_change.json
+++ b/test/scenarios/core/scenario_MoonModule_control_change.json
@@ -6,7 +6,7 @@
     "MultiplyModifier",
     "NoiseEffect"
   ],
-  "description": "Measure the cost of control changes on a running pipeline. Toggles MultiplyModifier's mirrorX/Y at different points and verifies each change is applied without freezing the render loop. Companion to the MoonModule control-change gate unit tests (unit_MoonModule_control_change_gate.cpp) — this is the live equivalent.",
+  "description": "Measure the cost of control changes on a running pipeline. Toggles MultiplyModifier's mirrorX/Y at different points and verifies each change is applied without freezing the render loop. Companion to the MoonModule control-change gate unit tests (unit_MoonModule_control_change_gate.cpp), this is the live equivalent.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -94,7 +94,7 @@
     },
     {
       "name": "baseline",
-      "description": "Set NoiseEffect.scale=4 and measure baseline FPS (mirror on). Effect controls don't rebuild the pipeline — slider stutter check.",
+      "description": "Set NoiseEffect.scale=4 and measure baseline FPS (mirror on). Effect controls don't rebuild the pipeline, slider stutter check.",
       "op": "set_control",
       "id": "Noise",
       "key": "scale",
@@ -276,7 +276,7 @@
     },
     {
       "name": "disable-mirrorX",
-      "description": "Disable mirrorX. Modifier control triggers a pipeline rebuild — measures the rebuilt path.",
+      "description": "Disable mirrorX. Modifier control triggers a pipeline rebuild, measures the rebuilt path.",
       "op": "set_control",
       "id": "Multiply",
       "key": "mirrorX",
@@ -458,7 +458,7 @@
     },
     {
       "name": "disable-mirrorY",
-      "description": "Disable mirrorY. Mirror is now fully off — should land on the no-LUT path.",
+      "description": "Disable mirrorY. Mirror is now fully off, should land on the no-LUT path.",
       "op": "set_control",
       "id": "Multiply",
       "key": "mirrorY",
@@ -648,7 +648,7 @@
     },
     {
       "name": "re-enable-mirrorY",
-      "description": "Re-enable mirrorY and measure — the heavy LUT path must recover (FPS within 50% of baseline) without staying degraded.",
+      "description": "Re-enable mirrorY and measure, the heavy LUT path must recover (FPS within 50% of baseline) without staying degraded.",
       "op": "set_control",
       "id": "Multiply",
       "key": "mirrorY",
diff --git a/test/scenarios/core/scenario_MqttModule_haDiscovery_toggle.json b/test/scenarios/core/scenario_MqttModule_haDiscovery_toggle.json
index bc21acc0..5de5a715 100644
--- a/test/scenarios/core/scenario_MqttModule_haDiscovery_toggle.json
+++ b/test/scenarios/core/scenario_MqttModule_haDiscovery_toggle.json
@@ -3,7 +3,7 @@
   "module": "MqttModule",
   "mode": "mutate",
   "live_only": true,
-  "description": "Toggle the Home-Assistant MQTT-discovery announce on and off and confirm the render loop survives every change. Pins the robustness + No-reboot-to-apply guarantees for the new haDiscovery control: turning it on publishes a retained homeassistant/light/<id>/config (and subscribes ha/set); turning it off retracts the config (empty retained payload) — both live over the already-connected socket on tick1s(), never a reconnect and never on the render task. No toggle, in any order, may crash, wedge the tick, or leak heap. Runs live only — discovery needs a real broker to publish to and MqttModule only exists on a board whose catalog entry includes it (the Shelly model does), so the in-process desktop runner SKIPs it; on a device it flips the control over HTTP. The bound checks FPS + heap stay within range across the on/off/on cycle (a discovery publish is a transient cost, not a permanent degradation), proving the toggle never leaves the network task or render loop degraded. (Background: the announce/retract are on the 1 Hz tick, off the hot path — see the MqttModule /// § Home Assistant MQTT Discovery.)",
+  "description": "Toggle the Home-Assistant MQTT-discovery announce on and off and confirm the render loop survives every change. Pins the robustness + No-reboot-to-apply guarantees for the new haDiscovery control: turning it on publishes a retained homeassistant/light/<id>/config (and subscribes ha/set); turning it off retracts the config (empty retained payload), both live over the already-connected socket on tick1s(), never a reconnect and never on the render task. No toggle, in any order, may crash, wedge the tick, or leak heap. Runs live only, discovery needs a real broker to publish to and MqttModule only exists on a board whose catalog entry includes it (the Shelly model does), so the in-process desktop runner SKIPs it; on a device it flips the control over HTTP. The bound checks FPS + heap stay within range across the on/off/on cycle (a discovery publish is a transient cost, not a permanent degradation), proving the toggle never leaves the network task or render loop degraded. (Background: the announce/retract are on the 1 Hz tick, off the hot path, see the MqttModule /// § Home Assistant MQTT Discovery.)",
   "reset": [
     {
       "name": "reset-haDiscovery",
@@ -16,7 +16,7 @@
   "steps": [
     {
       "name": "baseline-discovery-on",
-      "description": "haDiscovery on (default) — captures the baseline FPS/heap for the next two steps.",
+      "description": "haDiscovery on (default), captures the baseline FPS/heap for the next two steps.",
       "op": "set_control",
       "id": "Mqtt",
       "key": "haDiscovery",
@@ -27,7 +27,7 @@
           "tick_us": 100000,
           "free_heap": 90000,
           "set_by": "2026-07-06",
-          "reason": "initial contract — pending first hardware capture"
+          "reason": "initial contract, pending first hardware capture"
         }
       },
       "observed": {
@@ -100,7 +100,7 @@
     },
     {
       "name": "discovery-off",
-      "description": "haDiscovery off — the retract publish (empty retained config) goes out; FPS/heap must match or exceed the baseline.",
+      "description": "haDiscovery off, the retract publish (empty retained config) goes out; FPS/heap must match or exceed the baseline.",
       "op": "set_control",
       "id": "Mqtt",
       "key": "haDiscovery",
@@ -111,7 +111,7 @@
           "tick_us": 100000,
           "free_heap": 90000,
           "set_by": "2026-07-06",
-          "reason": "initial contract — pending first hardware capture"
+          "reason": "initial contract, pending first hardware capture"
         }
       },
       "observed": {
@@ -184,7 +184,7 @@
     },
     {
       "name": "discovery-on-again",
-      "description": "haDiscovery on again — re-announce + re-subscribe live. Bound: FPS must stay within 20% of the baseline (proves toggling doesn't leave the network task in a degraded state).",
+      "description": "haDiscovery on again, re-announce + re-subscribe live. Bound: FPS must stay within 20% of the baseline (proves toggling doesn't leave the network task in a degraded state).",
       "op": "set_control",
       "id": "Mqtt",
       "key": "haDiscovery",
@@ -195,7 +195,7 @@
           "tick_us": 100000,
           "free_heap": 90000,
           "set_by": "2026-07-06",
-          "reason": "initial contract — pending first hardware capture"
+          "reason": "initial contract, pending first hardware capture"
         }
       },
       "observed": {
diff --git a/test/scenarios/core/scenario_NetworkModule_eth_reconfigure.json b/test/scenarios/core/scenario_NetworkModule_eth_reconfigure.json
index 6805fe3e..abccc0cf 100644
--- a/test/scenarios/core/scenario_NetworkModule_eth_reconfigure.json
+++ b/test/scenarios/core/scenario_NetworkModule_eth_reconfigure.json
@@ -3,7 +3,7 @@
   "module": "NetworkModule",
   "mode": "mutate",
   "live_only": true,
-  "description": "Cycle the Ethernet PHY type (ethType: None/LAN8720/IP101/W5500) live on a running device and confirm the render loop survives every transition. Pins the robustness principle for the runtime Ethernet config: changing ethType reshapes the platform eth config (NetworkModule.syncEthLive → setEthConfig, with a live ethStop+ethInit on W5500), and no value or transition order may crash or wedge the tick. Runs live only — the eth controls exist only on hasEthernet ESP32 builds, so the in-process desktop runner SKIPs it; on a device it drives the controls over HTTP. On RMII boards (Olimex) the change saves + asks for restart (no hot re-init), so the live HTTP connection is undisturbed; the W5500 path hot-reinits but the SPI bus teardown keeps the netif alive enough to keep serving.",
+  "description": "Cycle the Ethernet PHY type (ethType: None/LAN8720/IP101/W5500) live on a running device and confirm the render loop survives every transition. Pins the robustness principle for the runtime Ethernet config: changing ethType reshapes the platform eth config (NetworkModule.syncEthLive → setEthConfig, with a live ethStop+ethInit on W5500), and no value or transition order may crash or wedge the tick. Runs live only, the eth controls exist only on hasEthernet ESP32 builds, so the in-process desktop runner SKIPs it; on a device it drives the controls over HTTP. On RMII boards (Olimex) the change saves + asks for restart (no hot re-init), so the live HTTP connection is undisturbed; the W5500 path hot-reinits but the SPI bus teardown keeps the netif alive enough to keep serving.",
   "reset": [
     {
       "name": "reset-grid-width",
@@ -30,7 +30,7 @@
   "steps": [
     {
       "name": "ethType-lan8720",
-      "description": "LAN8720 (RMII) — the classic-ESP32 default. Baseline: device renders with Ethernet RMII selected.",
+      "description": "LAN8720 (RMII), the classic-ESP32 default. Baseline: device renders with Ethernet RMII selected.",
       "op": "set_control",
       "id": "Network",
       "key": "ethType",
@@ -225,7 +225,7 @@
     },
     {
       "name": "ethType-ip101",
-      "description": "Switch to IP101 (RMII) live — the P4 PHY. Exercises the last dropdown value; render loop must survive.",
+      "description": "Switch to IP101 (RMII) live, the P4 PHY. Exercises the last dropdown value; render loop must survive.",
       "op": "set_control",
       "id": "Network",
       "key": "ethType",
@@ -290,7 +290,7 @@
     },
     {
       "name": "ethType-back-to-lan8720",
-      "description": "Return to LAN8720 — confirms the cycle is reversible and leaves the device in a sane state.",
+      "description": "Return to LAN8720, confirms the cycle is reversible and leaves the device in a sane state.",
       "op": "set_control",
       "id": "Network",
       "key": "ethType",
diff --git a/test/scenarios/core/scenario_NetworkModule_mdns_toggle.json b/test/scenarios/core/scenario_NetworkModule_mdns_toggle.json
index 1c51f198..1a89651b 100644
--- a/test/scenarios/core/scenario_NetworkModule_mdns_toggle.json
+++ b/test/scenarios/core/scenario_NetworkModule_mdns_toggle.json
@@ -16,7 +16,7 @@
   "steps": [
     {
       "name": "baseline-mdns-on",
-      "description": "mDNS on (default) — captures the baseline FPS for the next two steps.",
+      "description": "mDNS on (default), captures the baseline FPS for the next two steps.",
       "op": "set_control",
       "id": "Network",
       "key": "mDNS",
@@ -181,7 +181,7 @@
     },
     {
       "name": "mdns-off",
-      "description": "mDNS off — measured. Expected to match or exceed the baseline.",
+      "description": "mDNS off, measured. Expected to match or exceed the baseline.",
       "op": "set_control",
       "id": "Network",
       "key": "mDNS",
@@ -346,7 +346,7 @@
     },
     {
       "name": "mdns-on-again",
-      "description": "mDNS on again — measured with a bound: FPS must stay within 20% of the baseline (proves toggling doesn't leave the network task in a degraded state).",
+      "description": "mDNS on again, measured with a bound: FPS must stay within 20% of the baseline (proves toggling doesn't leave the network task in a degraded state).",
       "op": "set_control",
       "id": "Network",
       "key": "mDNS",
diff --git a/test/scenarios/light/scenario_Audio_mutation.json b/test/scenarios/light/scenario_Audio_mutation.json
index c3e7b447..45b2044d 100644
--- a/test/scenarios/light/scenario_Audio_mutation.json
+++ b/test/scenarios/light/scenario_Audio_mutation.json
@@ -48,7 +48,7 @@
     },
     {
       "name": "fix-layers",
-      "description": "Effects: the top-level container the Layer hangs under — the standard tree the live device boots with. A pinned single Layer (parent_id-less, props.layouts) is the older shape the host runner tolerates but the device's live reachability check does not.",
+      "description": "Effects: the top-level container the Layer hangs under, the standard tree the live device boots with. A pinned single Layer (parent_id-less, props.layouts) is the older shape the host runner tolerates but the device's live reachability check does not.",
       "op": "add_module",
       "id": "Effects",
       "type": "Effects",
@@ -490,7 +490,7 @@
     },
     {
       "name": "remove-audio-module",
-      "description": "Remove the mic while the consumer is STILL live. The consumer must fall back to AudioService::latestFrame()'s static silence — no dangling pointer, no crash. This is the robustness rule's hardest case for this pair.",
+      "description": "Remove the mic while the consumer is STILL live. The consumer must fall back to AudioService::latestFrame()'s static silence, no dangling pointer, no crash. This is the robustness rule's hardest case for this pair.",
       "op": "remove_module",
       "id": "Audio"
     },
@@ -585,7 +585,7 @@
     },
     {
       "name": "remove-audio-consumer",
-      "description": "Remove the orphaned consumer too — clean teardown, pipeline still live.",
+      "description": "Remove the orphaned consumer too, clean teardown, pipeline still live.",
       "op": "remove_module",
       "id": "Vol"
     },
diff --git a/test/scenarios/light/scenario_Driver_mutation.json b/test/scenarios/light/scenario_Driver_mutation.json
index 92dce12d..c0690939 100644
--- a/test/scenarios/light/scenario_Driver_mutation.json
+++ b/test/scenarios/light/scenario_Driver_mutation.json
@@ -9,7 +9,7 @@
     "Drivers",
     "PreviewDriver"
   ],
-  "description": "Add / remove output drivers under the Drivers container while the render pipeline runs, proving the robustness rule for driver deletion (the LED-driver delete path the product owner asked to cover; on the host RmtLed/I80Led/Parlio are inert via platform constants, so the portable NetworkSendDriver + PreviewDriver exercise the SAME generic add/remove lifecycle through the Scheduler that the hardware drivers use). Drivers are consumers of the Layer's output buffer, added/removed at runtime in any order. The checks assert the pipeline keeps RENDERING (buffer non-null, fps measurable) through each mutation: adding a second driver, removing it, and crucially removing a driver while the pipeline is still live (a driver teardown must release its resources without stranding the buffer or the other drivers — the same end-to-end Scheduler path the Audio producer/consumer scenario proves for peripherals, here for the output stage). Grid is 64x64 so the tick stays above the host microsecond clock at every step.",
+  "description": "Add / remove output drivers under the Drivers container while the render pipeline runs, proving the robustness rule for driver deletion (the LED-driver delete path the product owner asked to cover; on the host RmtLed/I80Led/Parlio are inert via platform constants, so the portable NetworkSendDriver + PreviewDriver exercise the SAME generic add/remove lifecycle through the Scheduler that the hardware drivers use). Drivers are consumers of the Layer's output buffer, added/removed at runtime in any order. The checks assert the pipeline keeps RENDERING (buffer non-null, fps measurable) through each mutation: adding a second driver, removing it, and crucially removing a driver while the pipeline is still live (a driver teardown must release its resources without stranding the buffer or the other drivers, the same end-to-end Scheduler path the Audio producer/consumer scenario proves for peripherals, here for the output stage). Grid is 64x64 so the tick stays above the host microsecond clock at every step.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -154,7 +154,7 @@
     },
     {
       "name": "add-second-driver",
-      "description": "Add a NetworkSendDriver beside the Preview driver — two consumers of the same Layer output buffer. The add must not disturb the running pipeline.",
+      "description": "Add a NetworkSendDriver beside the Preview driver, two consumers of the same Layer output buffer. The add must not disturb the running pipeline.",
       "op": "add_module",
       "id": "ArtNet",
       "type": "NetworkSendDriver",
@@ -251,7 +251,7 @@
     },
     {
       "name": "add-third-driver",
-      "description": "Add a second NetworkSendDriver — three drivers now share the Layer output buffer (Preview + two ArtNet). Stacking editable drivers before tearing them down in sequence.",
+      "description": "Add a second NetworkSendDriver, three drivers now share the Layer output buffer (Preview + two ArtNet). Stacking editable drivers before tearing them down in sequence.",
       "op": "add_module",
       "id": "ArtNet2",
       "type": "NetworkSendDriver",
@@ -348,7 +348,7 @@
     },
     {
       "name": "remove-first-added-driver",
-      "description": "Remove the FIRST-added NetworkSendDriver while the others (Preview + ArtNet2) are still live — delete a middle consumer, not the last. Its teardown must release its resources without stranding the buffer or the surviving drivers.",
+      "description": "Remove the FIRST-added NetworkSendDriver while the others (Preview + ArtNet2) are still live, delete a middle consumer, not the last. Its teardown must release its resources without stranding the buffer or the surviving drivers.",
       "op": "remove_module",
       "id": "ArtNet"
     },
@@ -443,13 +443,13 @@
     },
     {
       "name": "remove-second-added-driver",
-      "description": "Remove the remaining editable driver (ArtNet2) too — back to just the boot-wired Preview. Repeated driver teardown leaves no residue; the pipeline still renders. (PreviewDriver is userEditable=false, so it stays — drivers always have at least the wired output consumer; this mirrors the live API, which forbids deleting code-wired modules.)",
+      "description": "Remove the remaining editable driver (ArtNet2) too, back to just the boot-wired Preview. Repeated driver teardown leaves no residue; the pipeline still renders. (PreviewDriver is userEditable=false, so it stays, drivers always have at least the wired output consumer; this mirrors the live API, which forbids deleting code-wired modules.)",
       "op": "remove_module",
       "id": "ArtNet2"
     },
     {
       "name": "measure-back-to-one-driver",
-      "description": "Both added drivers gone, back to the single Preview baseline, still rendering — the add/remove cycle leaves the pipeline coherent.",
+      "description": "Both added drivers gone, back to the single Preview baseline, still rendering, the add/remove cycle leaves the pipeline coherent.",
       "op": "measure",
       "measure": true,
       "bounds": {
diff --git a/test/scenarios/light/scenario_Effects_composition.json b/test/scenarios/light/scenario_Effects_composition.json
index 4ec876fb..80d5bcf0 100644
--- a/test/scenarios/light/scenario_Effects_composition.json
+++ b/test/scenarios/light/scenario_Effects_composition.json
@@ -33,7 +33,7 @@
     },
     {
       "name": "add-layers-group",
-      "description": "Top-level Effects container — the multi-layer composition host.",
+      "description": "Top-level Effects container, the multi-layer composition host.",
       "op": "add_module",
       "id": "Effects",
       "type": "Effects",
@@ -43,7 +43,7 @@
     },
     {
       "name": "add-bottom-layer",
-      "description": "Bottom Layer (composited first — clears + overwrites the output buffer). RGB.",
+      "description": "Bottom Layer (composited first, clears + overwrites the output buffer). RGB.",
       "op": "add_module",
       "id": "LayerBottom",
       "type": "Layer",
@@ -62,7 +62,7 @@
     },
     {
       "name": "add-top-layer",
-      "description": "Top Layer (composited second — blends onto the bottom with its default blend). RGB.",
+      "description": "Top Layer (composited second, blends onto the bottom with its default blend). RGB.",
       "op": "add_module",
       "id": "LayerTop",
       "type": "Layer",
@@ -73,7 +73,7 @@
     },
     {
       "name": "add-top-effect",
-      "description": "Rainbow as the top layer's effect — composited over the Spiral base.",
+      "description": "Rainbow as the top layer's effect, composited over the Spiral base.",
       "op": "add_module",
       "id": "Rainbow",
       "type": "RainbowEffect",
diff --git a/test/scenarios/light/scenario_GridLayout_resize.json b/test/scenarios/light/scenario_GridLayout_resize.json
index ce0df001..51fe07d7 100644
--- a/test/scenarios/light/scenario_GridLayout_resize.json
+++ b/test/scenarios/light/scenario_GridLayout_resize.json
@@ -276,7 +276,7 @@
     },
     {
       "name": "shrink-to-128x64",
-      "description": "Shrink to 128x64. Measured: tick/heap captured so the runner reports the realloc behaviour against each target's contract. (The old relative-to-baseline FPS bound was removed — it compared against the runner's idle pre-scenario baseline, not the prior render step, so it false-failed on fast boards like the P4 where idle FPS dwarfs render FPS.)",
+      "description": "Shrink to 128x64. Measured: tick/heap captured so the runner reports the realloc behaviour against each target's contract. (The old relative-to-baseline FPS bound was removed, it compared against the runner's idle pre-scenario baseline, not the prior render step, so it false-failed on fast boards like the P4 where idle FPS dwarfs render FPS.)",
       "op": "set_control",
       "id": "Grid",
       "key": "height",
diff --git a/test/scenarios/light/scenario_Layer_base_pipeline.json b/test/scenarios/light/scenario_Layer_base_pipeline.json
index 6d097ceb..bb2ced5d 100644
--- a/test/scenarios/light/scenario_Layer_base_pipeline.json
+++ b/test/scenarios/light/scenario_Layer_base_pipeline.json
@@ -19,7 +19,7 @@
     },
     {
       "name": "add-grid",
-      "description": "Add a 128x128 GridLayout child to Layouts. Set explicitly (the module default is 16x16x1) so the tick is above the host's microsecond clock resolution — a 16x16 grid renders in <1us on desktop, flooring tick to 0.",
+      "description": "Add a 128x128 GridLayout child to Layouts. Set explicitly (the module default is 16x16x1) so the tick is above the host's microsecond clock resolution, a 16x16 grid renders in <1us on desktop, flooring tick to 0.",
       "op": "add_module",
       "id": "Grid",
       "type": "GridLayout",
diff --git a/test/scenarios/light/scenario_Layouts_mutation.json b/test/scenarios/light/scenario_Layouts_mutation.json
index cff36acb..e8f1c502 100644
--- a/test/scenarios/light/scenario_Layouts_mutation.json
+++ b/test/scenarios/light/scenario_Layouts_mutation.json
@@ -10,7 +10,7 @@
     "Drivers",
     "NetworkSendDriver"
   ],
-  "description": "Tree mutation on the Layouts container while the pipeline runs: add a second layout (multiple layouts under one Layouts), replace a layout with a different type, and remove a layout. The check is that each mutation leaves the pipeline RENDERING — Layer + Drivers re-wire via prepareTree and the buffer stays non-null and non-zero. Mirrors the HTTP add/replace/delete handlers; exercises the runner's add_module / replace_module / remove_module ops. NOTE: the Layer renders a dense bounding-box buffer sized by the layouts' coordinate EXTENT, not the summed light count — layouts that overlap in coordinate space share voxels (two 64x64 grids both occupy x,y in 0..63). There are no per-layout coordinate offsets, so multiple layouts share the same coordinate box; these steps assert liveness, not buffer-size arithmetic. Grids are 64x64 so the tick stays above the host's microsecond clock at every step.",
+  "description": "Tree mutation on the Layouts container while the pipeline runs: add a second layout (multiple layouts under one Layouts), replace a layout with a different type, and remove a layout. The check is that each mutation leaves the pipeline RENDERING, Layer + Drivers re-wire via prepareTree and the buffer stays non-null and non-zero. Mirrors the HTTP add/replace/delete handlers; exercises the runner's add_module / replace_module / remove_module ops. NOTE: the Layer renders a dense bounding-box buffer sized by the layouts' coordinate EXTENT, not the summed light count, layouts that overlap in coordinate space share voxels (two 64x64 grids both occupy x,y in 0..63). There are no per-layout coordinate offsets, so multiple layouts share the same coordinate box; these steps assert liveness, not buffer-size arithmetic. Grids are 64x64 so the tick stays above the host's microsecond clock at every step.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -183,7 +183,7 @@
     },
     {
       "name": "add-second-layout",
-      "description": "Add a SECOND layout (a 64x64 grid) under Layouts — two layouts now live under one container. prepareTree re-runs; the pipeline must still render. (Both grids share the 0..63 coordinate box, so the Layer buffer stays 64x64 — see the scenario NOTE.)",
+      "description": "Add a SECOND layout (a 64x64 grid) under Layouts, two layouts now live under one container. prepareTree re-runs; the pipeline must still render. (Both grids share the 0..63 coordinate box, so the Layer buffer stays 64x64, see the scenario NOTE.)",
       "op": "add_module",
       "id": "GridB",
       "type": "GridLayout",
@@ -318,7 +318,7 @@
     },
     {
       "name": "measure-after-replace",
-      "description": "Pipeline still renders after replacing a grid with a sphere (different layout type, same slot) — buffer re-wires without crashing.",
+      "description": "Pipeline still renders after replacing a grid with a sphere (different layout type, same slot), buffer re-wires without crashing.",
       "op": "measure",
       "measure": true,
       "bounds": {
@@ -434,7 +434,7 @@
     },
     {
       "name": "remove-second-layout",
-      "description": "Remove the sphere — back to a single grid layout. Layer/Drivers shrink their buffers via prepareTree.",
+      "description": "Remove the sphere, back to a single grid layout. Layer/Drivers shrink their buffers via prepareTree.",
       "op": "remove_module",
       "id": "GridB"
     },
diff --git a/test/scenarios/light/scenario_MoonLiveEffect_controls.json b/test/scenarios/light/scenario_MoonLiveEffect_controls.json
index 99af4282..064f6758 100644
--- a/test/scenarios/light/scenario_MoonLiveEffect_controls.json
+++ b/test/scenarios/light/scenario_MoonLiveEffect_controls.json
@@ -14,7 +14,7 @@
     "Drivers",
     "NetworkSendDriver"
   ],
-  "description": "Exercise MoonLive Stage-1 CONTROLS end-to-end as a wired module. A script declares a member and surfaces it (`addControl(\"speed\", speed, 0, 15)` in defineControls) and uses it (`setRGB(speed, ...)`); the engine surfaces the control, the binding creates a real MoonModule control of the widget its member's TYPE calls for, bound to the member's 4-byte slot in the live control-values arena. The scenario: add the effect with a control script (the control appears, renders), change the CONTROL value live (a slider move — must NOT recompile; the arena byte updates and the next tick reads it), edit the SOURCE to add a second control (recompile re-derives the set, existing slider value preserved by the stable-address grow-only arena), edit the source to remove a control (the orphaned value drops), push a broken script (compile fails, renders dark, status shows the diagnostic, no crash), recover, and remove + re-add (resource teardown + re-acquire). A crash in the LoadCtrl codegen, a dangling arena pointer across a recompile, or a value change that wrongly triggers a recompile all show up as a failed measure or a tick spike. The codegen + live-read contract is pinned by unit_moonlive_ir / unit_moonlive_compiler; this is the wired-module gate.",
+  "description": "Exercise MoonLive Stage-1 CONTROLS end-to-end as a wired module. A script declares a member and surfaces it (`addControl(\"speed\", speed, 0, 15)` in defineControls) and uses it (`setRGB(speed, ...)`); the engine surfaces the control, the binding creates a real MoonModule control of the widget its member's TYPE calls for, bound to the member's 4-byte slot in the live control-values arena. The scenario: add the effect with a control script (the control appears, renders), change the CONTROL value live (a slider move, must NOT recompile; the arena byte updates and the next tick reads it), edit the SOURCE to add a second control (recompile re-derives the set, existing slider value preserved by the stable-address grow-only arena), edit the source to remove a control (the orphaned value drops), push a broken script (compile fails, renders dark, status shows the diagnostic, no crash), recover, and remove + re-add (resource teardown + re-acquire). A crash in the LoadCtrl codegen, a dangling arena pointer across a recompile, or a value change that wrongly triggers a recompile all show up as a failed measure or a tick spike. The codegen + live-read contract is pinned by unit_moonlive_ir / unit_moonlive_compiler; this is the wired-module gate.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -227,7 +227,7 @@
     },
     {
       "name": "change-control-live",
-      "description": "Change the `speed` control value (a slider move). This must NOT recompile — affectsPrepare returns false for a scripted control; the arena byte updates and the next render tick reads it. Tick stays cheap (a recompile would spike it).",
+      "description": "Change the `speed` control value (a slider move). This must NOT recompile, affectsPrepare returns false for a scripted control; the arena byte updates and the next render tick reads it. Tick stays cheap (a recompile would spike it).",
       "op": "set_control",
       "id": "ML",
       "key": "speed",
@@ -476,7 +476,7 @@
     },
     {
       "name": "edit-source-broken",
-      "description": "Push a broken script. Compile fails, the previous code is freed, the effect renders dark and the parse error surfaces in status — no crash.",
+      "description": "Push a broken script. Compile fails, the previous code is freed, the effect renders dark and the parse error surfaces in status, no crash.",
       "op": "set_control",
       "id": "ML",
       "key": "script",
@@ -559,7 +559,7 @@
     },
     {
       "name": "edit-source-recover",
-      "description": "Recover with a valid control script — the effect compiles and renders again.",
+      "description": "Recover with a valid control script, the effect compiles and renders again.",
       "op": "set_control",
       "id": "ML",
       "key": "script",
@@ -635,13 +635,13 @@
     },
     {
       "name": "remove-control-effect",
-      "description": "Remove the effect — the engine releases its exec block AND its control arena (release).",
+      "description": "Remove the effect, the engine releases its exec block AND its control arena (release).",
       "op": "remove_module",
       "id": "ML"
     },
     {
       "name": "re-add-control-effect",
-      "description": "Re-add a fresh effect after the remove — exec memory + control arena re-acquired clean (it renders its default fill on add). Control re-acquisition itself is proven by the add-control-script step at the top: a freshly-added effect compiling a control source surfaces + seeds its control; construct-mode set_control can't apply a dynamically-added scripted control as the final asserted render, so the gate here is the bare re-add's liveness.",
+      "description": "Re-add a fresh effect after the remove, exec memory + control arena re-acquired clean (it renders its default fill on add). Control re-acquisition itself is proven by the add-control-script step at the top: a freshly-added effect compiling a control source surfaces + seeds its control; construct-mode set_control can't apply a dynamically-added scripted control as the final asserted render, so the gate here is the bare re-add's liveness.",
       "op": "add_module",
       "id": "ML2",
       "type": "MoonLiveEffect",
diff --git a/test/scenarios/light/scenario_MoonLiveEffect_livescript.json b/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
index 028ff537..ebdd2a5e 100644
--- a/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
+++ b/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
@@ -14,7 +14,7 @@
     "Drivers",
     "NetworkSendDriver"
   ],
-  "description": "Exercise a scripted MoonLiveEffect as a wired MoonModule end-to-end — the integration layer the unit tests can't reach. The effect compiles its `source` text to native code on-device and renders it into the Layer buffer each tick. Prepares its own canvas: Layout(Grid 16x16) + Layer + MoonLiveEffect, measures the default compile, then edits `source` live (a new fill color recompiles and keeps rendering), pushes a BROKEN script (compile fails, the previous code is freed, the effect renders dark and the parse error surfaces in status, no crash), recovers with a valid script, and finally removes + re-adds the effect (add/remove robustness in any order). A crash in the JIT/emit path, a failed recompile that wedges the tick, or a buffer overrun on an odd grid all show up as a failed measure. The compiler + emit golden bytes are pinned by unit_moonlive_compiler / unit_moonlive_fill; this is the live wired-module gate.",
+  "description": "Exercise a scripted MoonLiveEffect as a wired MoonModule end-to-end, the integration layer the unit tests can't reach. The effect compiles its `source` text to native code on-device and renders it into the Layer buffer each tick. Prepares its own canvas: Layout(Grid 16x16) + Layer + MoonLiveEffect, measures the default compile, then edits `source` live (a new fill color recompiles and keeps rendering), pushes a BROKEN script (compile fails, the previous code is freed, the effect renders dark and the parse error surfaces in status, no crash), recovers with a valid script, and finally removes + re-adds the effect (add/remove robustness in any order). A crash in the JIT/emit path, a failed recompile that wedges the tick, or a buffer overrun on an odd grid all show up as a failed measure. The compiler + emit golden bytes are pinned by unit_moonlive_compiler / unit_moonlive_fill; this is the live wired-module gate.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -307,7 +307,7 @@
     },
     {
       "name": "edit-source-broken",
-      "description": "Push a script that fails to parse. The compile fails, the engine reports the diagnostic in the module status and renders dark, but the device keeps running (robust, no reboot) — the script-editor failure path. The measure passes because the pipeline still ticks.",
+      "description": "Push a script that fails to parse. The compile fails, the engine reports the diagnostic in the module status and renders dark, but the device keeps running (robust, no reboot), the script-editor failure path. The measure passes because the pipeline still ticks.",
       "op": "set_control",
       "id": "ML",
       "key": "script",
@@ -417,7 +417,7 @@
     },
     {
       "name": "edit-source-recover",
-      "description": "Push a valid script again. The engine recompiles cleanly and rendering resumes — a broken edit is fully recoverable.",
+      "description": "Push a valid script again. The engine recompiles cleanly and rendering resumes, a broken edit is fully recoverable.",
       "op": "set_control",
       "id": "ML",
       "key": "script",
@@ -520,7 +520,7 @@
     },
     {
       "name": "shrink-grid-1x1",
-      "description": "Resize the canvas to 1x1 while the scripted effect renders — the smallest non-empty grid. The native fill loops over a single light; the run guards (non-null buffer, cpl>=3) keep it in-bounds. Pins the 'runs at every grid size' hard rule for the JIT'd routine.",
+      "description": "Resize the canvas to 1x1 while the scripted effect renders, the smallest non-empty grid. The native fill loops over a single light; the run guards (non-null buffer, cpl>=3) keep it in-bounds. Pins the 'runs at every grid size' hard rule for the JIT'd routine.",
       "op": "set_control",
       "id": "Grid",
       "key": "width",
@@ -893,7 +893,7 @@
     },
     {
       "name": "re-add-moonlive",
-      "description": "Re-add a MoonLiveEffect after removal — the exec memory is re-acquired fresh, no leak, no stale pointer. The scripted effect renders again.",
+      "description": "Re-add a MoonLiveEffect after removal, the exec memory is re-acquired fresh, no leak, no stale pointer. The scripted effect renders again.",
       "op": "add_module",
       "id": "ML2",
       "type": "MoonLiveEffect",
diff --git a/test/scenarios/light/scenario_MoonLive_pipeline.json b/test/scenarios/light/scenario_MoonLive_pipeline.json
index ebf57196..2f2a1fa2 100644
--- a/test/scenarios/light/scenario_MoonLive_pipeline.json
+++ b/test/scenarios/light/scenario_MoonLive_pipeline.json
@@ -55,7 +55,7 @@
     },
     {
       "name": "select-default-layout",
-      "description": "Point the scripted layout at the SHIPPED grid.mll — the fixture a user starts from. A shipped file, not a staged one: the live runner skips the fixture, so anything written there is absent on a real device.",
+      "description": "Point the scripted layout at the SHIPPED grid.mll, the fixture a user starts from. A shipped file, not a staged one: the live runner skips the fixture, so anything written there is absent on a real device.",
       "op": "set_control",
       "id": "MLL",
       "key": "script",
diff --git a/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json b/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
index f1687cdc..e59fd3e7 100644
--- a/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
+++ b/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
@@ -49,7 +49,7 @@
     },
     {
       "name": "add-mirror",
-      "description": "Add MultiplyModifier — triggers LUT and driver-buffer allocation.",
+      "description": "Add MultiplyModifier, triggers LUT and driver-buffer allocation.",
       "op": "add_module",
       "id": "Multiply",
       "type": "MultiplyModifier",
diff --git a/test/scenarios/light/scenario_modifier_chain.json b/test/scenarios/light/scenario_modifier_chain.json
index 4fd587ad..c4c39498 100644
--- a/test/scenarios/light/scenario_modifier_chain.json
+++ b/test/scenarios/light/scenario_modifier_chain.json
@@ -13,7 +13,7 @@
     "Drivers",
     "PreviewDriver"
   ],
-  "description": "Stack TWO modifiers on one Layer (Region then Multiply) and verify the chain composes live end-to-end — the capability the old single-modifier engine couldn't do. Prepares its own canvas: Layout(Grid 32x32) + Layer + NoiseEffect + Region(0..50) + Multiply(2x), measures the composite, then adds a third (Checkerboard mask) and measures again, then removes the middle modifier and measures — exercising add/remove on a multi-modifier chain. A broken fold (null buffer, wrong light count, crash on a disabled/removed stage) shows up as a failed measure. The fold composition + order semantics are pinned by unit_Layer_modifier_chain; this is the live end-to-end gate.",
+  "description": "Stack TWO modifiers on one Layer (Region then Multiply) and verify the chain composes live end-to-end, the capability the old single-modifier engine couldn't do. Prepares its own canvas: Layout(Grid 32x32) + Layer + NoiseEffect + Region(0..50) + Multiply(2x), measures the composite, then adds a third (Checkerboard mask) and measures again, then removes the middle modifier and measures, exercising add/remove on a multi-modifier chain. A broken fold (null buffer, wrong light count, crash on a disabled/removed stage) shows up as a failed measure. The fold composition + order semantics are pinned by unit_Layer_modifier_chain; this is the live end-to-end gate.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -153,7 +153,7 @@
     },
     {
       "name": "add-mask",
-      "description": "Add a third modifier (Checkerboard mask) on top of the chain — a 3-deep fold. Measure that the deeper chain still renders.",
+      "description": "Add a third modifier (Checkerboard mask) on top of the chain, a 3-deep fold. Measure that the deeper chain still renders.",
       "op": "add_module",
       "id": "MASK",
       "type": "CheckerboardModifier",
@@ -213,7 +213,7 @@
     },
     {
       "name": "remove-middle",
-      "description": "Remove the middle modifier (Multiply) — the chain re-folds with Region then Checkerboard, no stale state. Measure.",
+      "description": "Remove the middle modifier (Multiply), the chain re-folds with Region then Checkerboard, no stale state. Measure.",
       "op": "remove_module",
       "id": "MULT",
       "measure": true,
@@ -244,7 +244,7 @@
     },
     {
       "name": "add-live-rotate",
-      "description": "Add a DYNAMIC Rotate on top of the static chain — its modifyLive runs the per-frame remap pass over the composed buffer. Verifies a static chain + a live modifier coexist (the buffer is remapped each frame on top of the baked Region/Checkerboard mapping) without a crash or null buffer.",
+      "description": "Add a DYNAMIC Rotate on top of the static chain, its modifyLive runs the per-frame remap pass over the composed buffer. Verifies a static chain + a live modifier coexist (the buffer is remapped each frame on top of the baked Region/Checkerboard mapping) without a crash or null buffer.",
       "op": "add_module",
       "id": "ROT",
       "type": "RotateModifier",
diff --git a/test/scenarios/light/scenario_modifier_swap.json b/test/scenarios/light/scenario_modifier_swap.json
index 3e0e42c4..6d4cffe3 100644
--- a/test/scenarios/light/scenario_modifier_swap.json
+++ b/test/scenarios/light/scenario_modifier_swap.json
@@ -147,7 +147,7 @@
       "name": "multiply-1",
       "op": "measure",
       "measure": true,
-      "description": "Multiply modifier active — pipeline live, LUT folds the grid.",
+      "description": "Multiply modifier active, pipeline live, LUT folds the grid.",
       "observed": {
         "desktop-macos": {
           "tick_us": {
@@ -291,7 +291,7 @@
       "name": "checkerboard",
       "op": "measure",
       "measure": true,
-      "description": "Checkerboard modifier active — masks half the lights; pipeline stays live (driver buffer non-null).",
+      "description": "Checkerboard modifier active, masks half the lights; pipeline stays live (driver buffer non-null).",
       "observed": {
         "desktop-macos": {
           "tick_us": {
@@ -435,7 +435,7 @@
       "name": "multiply-2",
       "op": "measure",
       "measure": true,
-      "description": "Back to Multiply — replace round-trips cleanly, pipeline live again.",
+      "description": "Back to Multiply, replace round-trips cleanly, pipeline live again.",
       "observed": {
         "desktop-macos": {
           "tick_us": {
diff --git a/test/scenarios/light/scenario_perf_full.json b/test/scenarios/light/scenario_perf_full.json
index d17aa7ee..d8784adc 100644
--- a/test/scenarios/light/scenario_perf_full.json
+++ b/test/scenarios/light/scenario_perf_full.json
@@ -14,7 +14,7 @@
     "SpiralEffect",
     "NoiseEffect"
   ],
-  "description": "Comprehensive incremental performance check (the SLOW, on-device companion to scenario_perf_light). Mutate mode + canvas-preparing: clear_children whatever the device already had (pre-wired apparatus like PreviewDriver/Board survives — clear_children only drops user-editable children), rebuild a known minimal tree, then add one subsystem at a time — audio, device discovery, a modifier, then EVERY output driver this board has (each optional + capped to 64 output LEDs so its per-frame cost is comparable, not its transmit-all-16K time), then a network driver — measuring the tick/heap delta after each so each subsystem's cost is isolated. Then sweep the grid 16²→32²→64²→128² (16K) for both a LIGHT effect (Spiral) and a HEAVY one (Noise) to bracket the compute range across sizes. LED drivers are platform-gated (RMT on classic/S3, LCD on S3, Parlio on P4; none on desktop) so each driver step is optional:true and skipped where absent — the all-drivers comparison is assembled across boards (S3 gives RMT vs LCD, P4 gives RMT vs Parlio). Subsumes the old scenario_Layer_buildup (incremental module cost), scenario_GridLayout_grid_sizes (grid sweep), and scenario_AllEffects_grid_sizes (per-effect size sweep, here reduced to a light/heavy bracket). Runs minutes on a device; not a per-commit gate.",
+  "description": "Comprehensive incremental performance check (the SLOW, on-device companion to scenario_perf_light). Mutate mode + canvas-preparing: clear_children whatever the device already had (pre-wired apparatus like PreviewDriver/Board survives, clear_children only drops user-editable children), rebuild a known minimal tree, then add one subsystem at a time, audio, device discovery, a modifier, then EVERY output driver this board has (each optional + capped to 64 output LEDs so its per-frame cost is comparable, not its transmit-all-16K time), then a network driver, measuring the tick/heap delta after each so each subsystem's cost is isolated. Then sweep the grid 16²→32²→64²→128² (16K) for both a LIGHT effect (Spiral) and a HEAVY one (Noise) to bracket the compute range across sizes. LED drivers are platform-gated (RMT on classic/S3, LCD on S3, Parlio on P4; none on desktop) so each driver step is optional:true and skipped where absent, the all-drivers comparison is assembled across boards (S3 gives RMT vs LCD, P4 gives RMT vs Parlio). Subsumes the old scenario_Layer_buildup (incremental module cost), scenario_GridLayout_grid_sizes (grid sweep), and scenario_AllEffects_grid_sizes (per-effect size sweep, here reduced to a light/heavy bracket). Runs minutes on a device; not a per-commit gate.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -673,7 +673,7 @@
     },
     {
       "name": "add-network-driver",
-      "description": "+NetworkSendDriver (ArtNet/DDP — works on every platform). Diff = the network output cost (capped by the 16² grid here).",
+      "description": "+NetworkSendDriver (ArtNet/DDP, works on every platform). Diff = the network output cost (capped by the 16² grid here).",
       "op": "add_module",
       "id": "Net",
       "type": "NetworkSendDriver",
@@ -796,7 +796,7 @@
     },
     {
       "name": "add-rmt-driver",
-      "description": "+RmtLedDriver capped to 64 output LEDs (one pin, ledsPerPin=64). Optional — classic + S3. Diff = the RMT per-frame cost at a fixed 64-LED output.",
+      "description": "+RmtLedDriver capped to 64 output LEDs (one pin, ledsPerPin=64). Optional, classic + S3. Diff = the RMT per-frame cost at a fixed 64-LED output.",
       "op": "add_module",
       "id": "Rmt",
       "type": "RmtLedDriver",
@@ -926,7 +926,7 @@
     },
     {
       "name": "add-i80-driver",
-      "description": "+ParallelLedDriver (peripheral=i80) capped to 64 LEDs on lane 0 (i80 needs all 8 data pins; unused lanes get 0 LEDs). S3 ONLY — the pins below are the S3's clean set. The i80 peripheral also runs on the classic ESP32 (I2S backend) and the P4 (LCD_CAM), but NO 8-pin set is safe on all three (only GPIO 21 is common: the classic's 6-11 are the SPI flash bus, the P4 exposes a different range), so one shared step cannot measure them. On a non-S3 target the driver fails to init and the measure below is meaningless — the observed blocks are therefore S3/desktop only. Diff = the i80 per-frame cost.",
+      "description": "+ParallelLedDriver (peripheral=i80) capped to 64 LEDs on lane 0 (i80 needs all 8 data pins; unused lanes get 0 LEDs). S3 ONLY, the pins below are the S3's clean set. The i80 peripheral also runs on the classic ESP32 (I2S backend) and the P4 (LCD_CAM), but NO 8-pin set is safe on all three (only GPIO 21 is common: the classic's 6-11 are the SPI flash bus, the P4 exposes a different range), so one shared step cannot measure them. On a non-S3 target the driver fails to init and the measure below is meaningless, the observed blocks are therefore S3/desktop only. Diff = the i80 per-frame cost.",
       "op": "add_module",
       "id": "MultiPin",
       "type": "ParallelLedDriver",
@@ -1023,7 +1023,7 @@
           "last_updated": "2026-09-01"
         }
       },
-      "description": "The i80 per-frame cost. Only meaningful where add-i80-driver's pins are valid (S3; desktop is inert but exercises the code path) — the classic/P4 blocks were removed: they recorded a driver that had failed to init, so the tiny tick was the bail-out path, not an encode."
+      "description": "The i80 per-frame cost. Only meaningful where add-i80-driver's pins are valid (S3; desktop is inert but exercises the code path), the classic/P4 blocks were removed: they recorded a driver that had failed to init, so the tiny tick was the bail-out path, not an encode."
     },
     {
       "name": "remove-i80-driver",
@@ -1033,7 +1033,7 @@
     },
     {
       "name": "add-parlio-driver",
-      "description": "+ParallelLedDriver (peripheral=Parlio) capped to 64 LEDs on lane 0. Optional — P4 only. Diff = the Parlio per-frame cost.",
+      "description": "+ParallelLedDriver (peripheral=Parlio) capped to 64 LEDs on lane 0. Optional, P4 only. Diff = the Parlio per-frame cost.",
       "op": "add_module",
       "id": "Parlio",
       "type": "ParallelLedDriver",
@@ -1645,7 +1645,7 @@
     },
     {
       "name": "swap-heavy",
-      "description": "Swap to the HEAVY effect (Noise) and repeat the sweep — the upper bracket of per-pixel compute.",
+      "description": "Swap to the HEAVY effect (Noise) and repeat the sweep, the upper bracket of per-pixel compute.",
       "op": "replace_module",
       "id": "FX",
       "type": "NoiseEffect",
@@ -2149,7 +2149,7 @@
     },
     {
       "name": "add-modifier-for-sweep",
-      "description": "Re-add the MultiplyModifier on the HEAVY effect and sweep grid sizes — the diff vs the matching measure-heavy-N step (no modifier) is the modifier's PER-FRAME cost at each grid, answering whether the ~180µs seen at 16² scales with pixel count or is fixed overhead.",
+      "description": "Re-add the MultiplyModifier on the HEAVY effect and sweep grid sizes, the diff vs the matching measure-heavy-N step (no modifier) is the modifier's PER-FRAME cost at each grid, answering whether the ~180µs seen at 16² scales with pixel count or is fixed overhead.",
       "op": "add_module",
       "id": "Mod",
       "type": "MultiplyModifier",
diff --git a/test/scenarios/light/scenario_perf_light.json b/test/scenarios/light/scenario_perf_light.json
index b43518cc..5b7c7d5c 100644
--- a/test/scenarios/light/scenario_perf_light.json
+++ b/test/scenarios/light/scenario_perf_light.json
@@ -11,7 +11,7 @@
     "NoiseEffect",
     "MultiplyModifier"
   ],
-  "description": "Fast incremental performance check: start from the bare minimum render pipeline and add one thing at a time, measuring the tick/heap delta each step, so a regression shows up as a per-step jump. The LIGHT companion to scenario_perf_full — it stays small (≤64²) and driver-free so it runs in seconds. Mutate mode + canvas-preparing: the steps clear_children whatever Layouts/Effects/Drivers the device already had (the pre-wired apparatus like PreviewDriver/Board survives — clear_children only drops user-editable children) and rebuild a known tree, so it runs from any starting state and always measures the same minimal pipeline. Order: (1) minimal = Grid(16²)+Layer+a LIGHT effect (Spiral, a light effect), no modifier/driver/audio/discovery; (2) +MultiplyModifier (adds the mapping LUT — the heavy memory path); (3) +PreviewDriver; (4) swap to a HEAVY effect (Noise) to bracket the compute range; (5) grid 16²→32²→64² to show the size scaling. Full 128²/16K sweep, real LED/network drivers, audio+discovery cost: see scenario_perf_full.",
+  "description": "Fast incremental performance check: start from the bare minimum render pipeline and add one thing at a time, measuring the tick/heap delta each step, so a regression shows up as a per-step jump. The LIGHT companion to scenario_perf_full, it stays small (≤64²) and driver-free so it runs in seconds. Mutate mode + canvas-preparing: the steps clear_children whatever Layouts/Effects/Drivers the device already had (the pre-wired apparatus like PreviewDriver/Board survives, clear_children only drops user-editable children) and rebuild a known tree, so it runs from any starting state and always measures the same minimal pipeline. Order: (1) minimal = Grid(16²)+Layer+a LIGHT effect (Spiral, a light effect), no modifier/driver/audio/discovery; (2) +MultiplyModifier (adds the mapping LUT, the heavy memory path); (3) +PreviewDriver; (4) swap to a HEAVY effect (Noise) to bracket the compute range; (5) grid 16²→32²→64² to show the size scaling. Full 128²/16K sweep, real LED/network drivers, audio+discovery cost: see scenario_perf_full.",
   "fixture": [
     {
       "name": "fix-layouts",
@@ -32,7 +32,7 @@
   "steps": [
     {
       "name": "disable-audio",
-      "description": "Quiet I2S sampling so it can't pollute the tick (optional — device only). Targets the mic by its runtime name 'Audio' (displayNameFor strips the 'Service' suffix; the catalog adds it with id 'Audio').",
+      "description": "Quiet I2S sampling so it can't pollute the tick (optional, device only). Targets the mic by its runtime name 'Audio' (displayNameFor strips the 'Service' suffix; the catalog adds it with id 'Audio').",
       "op": "set_control",
       "id": "Audio",
       "key": "enabled",
@@ -41,7 +41,7 @@
     },
     {
       "name": "disable-devices",
-      "description": "Stop the blocking HTTP discovery sweep (optional — device only).",
+      "description": "Stop the blocking HTTP discovery sweep (optional, device only).",
       "op": "set_control",
       "id": "Devices",
       "key": "enabled",
@@ -50,7 +50,7 @@
     },
     {
       "name": "clear-layers",
-      "description": "Drop whatever effects/modifiers the device had — start clean.",
+      "description": "Drop whatever effects/modifiers the device had, start clean.",
       "op": "clear_children",
       "id": "Effects"
     },
@@ -325,7 +325,7 @@
     },
     {
       "name": "measure-with-preview",
-      "description": "PreviewDriver is the pre-wired apparatus — it survives clear_children and is already attached, so the measures above already include it (no add step needed; adding a second Preview is rejected). This is a stable repeat of the effect+modifier config for run-to-run variance.",
+      "description": "PreviewDriver is the pre-wired apparatus, it survives clear_children and is already attached, so the measures above already include it (no add step needed; adding a second Preview is rejected). This is a stable repeat of the effect+modifier config for run-to-run variance.",
       "op": "measure",
       "measure": true,
       "observed": {
@@ -436,7 +436,7 @@
     },
     {
       "name": "swap-heavy-fx",
-      "description": "Swap the light effect for a HEAVY one (Noise — simplex per pixel) to bracket the compute range at the same 16².",
+      "description": "Swap the light effect for a HEAVY one (Noise, simplex per pixel) to bracket the compute range at the same 16².",
       "op": "replace_module",
       "id": "FX",
       "type": "NoiseEffect",
diff --git a/test/scenarios/light/scenario_peripheral_switch.json b/test/scenarios/light/scenario_peripheral_switch.json
index 416a5aa7..fc482833 100644
--- a/test/scenarios/light/scenario_peripheral_switch.json
+++ b/test/scenarios/light/scenario_peripheral_switch.json
@@ -250,7 +250,7 @@
     },
     {
       "name": "i80-doublebuffer-on",
-      "description": "i80 supports the async double-buffer (esp_lcd has a real transaction queue) — turn it on.",
+      "description": "i80 supports the async double-buffer (esp_lcd has a real transaction queue), turn it on.",
       "op": "set_control",
       "id": "ParallelLed",
       "key": "doubleBuffer",
@@ -492,7 +492,7 @@
     },
     {
       "name": "moonI80-doublebuffer-off",
-      "description": "Explicitly turn doubleBuffer off on MoonI80 (a no-op behaviorally — it was already forced single — but exercises the control write on the hidden control).",
+      "description": "Explicitly turn doubleBuffer off on MoonI80 (a no-op behaviorally, it was already forced single, but exercises the control write on the hidden control).",
       "op": "set_control",
       "id": "ParallelLed",
       "key": "doubleBuffer",
@@ -750,7 +750,7 @@
     },
     {
       "name": "switch-back-i80",
-      "description": "Round-trip back to i80 — the driver must return to a clean, running state after the whole cycle.",
+      "description": "Round-trip back to i80, the driver must return to a clean, running state after the whole cycle.",
       "op": "set_control",
       "id": "ParallelLed",
       "key": "peripheral",
diff --git a/test/unit/light/unit_PreviewDriver.cpp b/test/unit/light/unit_PreviewDriver.cpp
index a2aea90b..927e79f3 100644
--- a/test/unit/light/unit_PreviewDriver.cpp
+++ b/test/unit/light/unit_PreviewDriver.cpp
@@ -86,8 +86,12 @@ struct CaptureBroadcaster : mm::BinaryBroadcaster {
     int coordCount() const { return lastCoord.size() >= 5 ? static_cast<int>(u32le(lastCoord, 1)) : -1; }
     int frameCount() const { return lastFrame.size() >= 5 ? static_cast<int>(u32le(lastFrame, 1)) : -1; }
     int coordStride() const { return lastCoord.size() >= 11 ? lastCoord[8] | (lastCoord[9] << 8) : -1; }
+    // The color frame's own stride, which the link paces independently of the table's.
+    int frameStride() const { return lastFrame.size() >= 9 ? lastFrame[5] | (lastFrame[6] << 8) : -1; }
     int coordEpoch() const { return lastCoord.size() >= 11 ? lastCoord[10] : -1; }
     int frameEpoch() const { return lastFrame.size() >= 9 ? lastFrame[7] : -1; }
+    // The lights the BODY carries, which must equal what the header advertises or the browser misreads every frame.
+    int frameBodyLights() const { return lastFrame.size() >= 9 ? static_cast<int>((lastFrame.size() - 9) / 3) : -1; }
     int frameDrops() const { return lastFrame.size() >= 9 ? lastFrame[8] : -1; }
 
 private:
@@ -211,7 +215,7 @@ TEST_CASE("PreviewDriver small grid sends all lights exactly") {
     CHECK(rig.cap.coordStride() == 1);
 }
 
-// A large layout is SPATIALLY downsampled (a regular per-axis lattice, not every-Nth-flat-index) so the payload fits the send-buffer cap without the diagonal moiré that linear stride produced on a grid whose width didn't divide the stride. The wire "stride" field carries the per-axis lattice/downscale factor (color k still maps 1:1 to coord k).
+// A large layout is SPATIALLY downsampled on a regular per-axis lattice, and the COLOR frame carries that stride while the coordinate table stays whole, so the browser maps each color onto its block.
 TEST_CASE("PreviewDriver downsamples on a regular spatial lattice when a client asks coarser") {
     // There is NO display cap: a host build (unlimited memory) serves any layout at full detail, and coarseness exists only as a client REQUEST. Ask for 1/2 and pin the lattice geometry. The extent (199) is ≤255/axis, so positions are sent at EXACT integer grid coordinates (no byte-scaling rounding), letting the regularity check below compare true lattice positions.
     mm::GridLayout g;
@@ -222,10 +226,12 @@ TEST_CASE("PreviewDriver downsamples on a regular spatial lattice when a client
     mm::platform::setTestNowMs(0);
     rig.produce();
 
-    CHECK(rig.cap.coordStride() == 2);            // served exactly as asked
-    CHECK(rig.cap.coordCount() == 100 * 100);     // ceil(200/2) per axis
+    // The FRAME carries the requested stride; the table stays full so the browser can upsample.
+    CHECK(rig.cap.frameStride() == 2);            // served exactly as asked
+    CHECK(rig.cap.frameCount() == 100 * 100);     // ceil(200/2) per axis
+    CHECK(rig.cap.coordStride() == 1);            // geometry is sent once, at full detail
+    CHECK(rig.cap.coordCount() == 200 * 200);     // every light, whatever the link asks
     CHECK(rig.cap.coordCount() > 0);
-    CHECK(rig.cap.coordCount() == rig.cap.frameCount());  // table + RGB agree (lockstep)
 
     // Regular lattice check: every sent X coordinate is a multiple of the same step, and so is every Y, i.e. the kept points sit on a grid, with NO per-row column drift (the diagonal-streak bug). Read the packed u8 positions back from the coord message.
     const auto& cd = rig.cap.lastCoord;
@@ -387,9 +393,10 @@ TEST_CASE("PreviewDriver dense downsample packs colors by closed-form index, in
     mm::platform::setTestNowMs(2000); rig.preview->tick();
     mm::platform::setTestNowMs(0);
     rig.produce();                                     // ask for + receive the table (pull model)
-    const int s = rig.cap.coordStride();
+    // The COLOR frame strides; the table stays whole for the browser to upsample onto.
+    const int s = rig.cap.frameStride();
     REQUIRE(s == 3);                                   // served exactly as asked
-    const int kept = rig.cap.coordCount();
+    const int kept = rig.cap.frameCount();
     REQUIRE(kept == (width + s - 1) / s);              // ceil(width/s) — closed-form count
 
     // Paint the 2nd kept column (x = s) bright green; the rest black.
@@ -511,3 +518,38 @@ TEST_CASE("A rig with no moving heads sends no aim message at all") {
     CHECK_FALSE(rig.preview->sendAim());                   // declines before doing any work
     CHECK(rig.cap.lastFrame.empty());                      // and nothing reached the wire
 }
+
+// Geometry is sent once per epoch at full resolution and colors carry the link's stride, so a slow link shows the whole layout at less detail rather than fewer lights.
+TEST_CASE("PreviewDriver sends the table whole and strides only the colors") {
+    mm::GridLayout g;
+    g.width = 16; g.height = 16; g.depth = 1;
+    PreviewRig rig(&g);
+    rig.cap.ask(4);                                   // a slow link: 1/4
+    mm::platform::setTestNowMs(2000); rig.preview->tick();
+    mm::platform::setTestNowMs(0);
+    rig.produce();
+
+    CHECK(rig.cap.coordStride() == 1);                // every light's position, once
+    CHECK(rig.cap.coordCount() == 16 * 16);
+    CHECK(rig.cap.frameStride() == 4);                // the link's pacing, on the colors alone
+    CHECK(rig.cap.frameCount() == 4 * 4);             // ceil(16/4) per axis
+    // The whole point: the browser can still draw every light, because it has every position.
+    CHECK(rig.cap.coordCount() > rig.cap.frameCount());
+}
+
+// The SPARSE path at a coarse stride: a mapped layout gathers through the cached index, built at the TABLE's stride, so a coarser frame must filter it or the body and header disagree.
+TEST_CASE("PreviewDriver strides a mapped layout's colors without re-walking the layout") {
+    mm::SphereLayout sph;
+    sph.radius = 8;
+    PreviewRig rig(&sph);
+    rig.cap.ask(2);                                   // a slow link, on a layout with no dense grid
+    mm::platform::setTestNowMs(2000); rig.preview->tick();
+    mm::platform::setTestNowMs(0);
+    rig.produce();
+
+    CHECK(rig.cap.coordStride() == 1);                // the table stays whole
+    CHECK(rig.cap.frameStride() == 2);                // the frame carries the link's stride
+    // The header's count IS the body's: a mismatch is the bug this pins.
+    CHECK(rig.cap.frameCount() < rig.cap.coordCount());
+    CHECK(rig.cap.frameBodyLights() == rig.cap.frameCount());
+}

From eaa904dc5f4b0aaede67b4e6a88282fe73376b50 Mon Sep 17 00:00:00 2001
From: ewowi <ewowi@icloud.com>
Date: Mon, 21 Sep 2026 23:04:32 +0200
Subject: [PATCH 4/5] Flash an S31 from the browser, and keep every light its
 stride keeps
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The web installer now flashes every chip projectMM ships, the S31 included, on esptool-js 0.7.0. Fixes a preview cache that dropped lights whose coordinates shared a large factor, and adds a rule that each change leaves the documentation warnings in its own files lower than it found them.

KPI: 16384lights | Desktop:1924KB | tick:121/91/391/2/6/125/282/18/1355/3/274/66/243/17/22/6/1682/125/21/5/153/41/5us(FPS:8264/10989/2557/500000/166666/8000/3546/55555/738/333333/3649/15151/4115/58823/45454/166666/594/8000/47619/200000/6535/24390/200000) | src:270(68900) | test:206(45223) | lizard:267w

**Web installer**
- esptool-js 0.5.7 → 0.7.0, the first release carrying an ESP32-S31 target and chip-id detection (GET_SECURITY_INFO). The S31's ROM magic collides with the classic ESP32's, so a magic-only table would have flashed the wrong stub; identifying by chip id is what makes it safe.
- `writeFlash` takes a `Uint8Array`: 0.7.0 throws on the binary string the old API required, and passing one corrupted the image rather than refusing it.
- `WEB_FLASH_UNSUPPORTED_CHIPS` is empty. The guard stays for the next chip esptool-js does not know yet.
- Verified by browser-flashing four chips: P4, S31, classic Olimex and S3. The P4 also settles the deterministic 0.6.0 flash abort that held the pin at 0.5.7 since June.

**Light domain**
- The preview's kept-light cache stores each light's coordinates rather than one "coarsest divisor". That divisor answers the stride test only while it IS the gcd, and the loop capped at 64: a light whose axes shared a larger factor was dropped from a frame every axis divides by. A 51,554-light sphere now sends 6510 lights at stride 2 where it sent 6507.
- The driver's heap readout counts that array.

**UI**
- A preview point whose block holds no kept light draws dark instead of borrowing the first light's color, which is the normal case on a sparse layout. A color index outside the frame is skipped rather than read past the body.
- The stats tooltip describes the gesture it has: hold for fps, release for time.

**Core**
- `NetworkModule` keeps its pin-seeding comment honest about when the seed runs.

**Scripts/MoonDeck**
- `check_docgen` keys markdown hard-wrap findings on the docs-relative path the catalog findings use, so one page cannot appear twice under two spellings.

**Docs/CI**
- CLAUDE.md gains the warnings rule: holding the ratchet is the floor, and a change leaves the warnings in files it already touches lower than it found them. Applied here, 3135 → 3131, with no rule paying for another.
- Four British spellings and three em-dashes, plus a stray comma an earlier em-dash sweep left in a MIGRATING marker cell.
- `building.md` no longer tells a reader the S31 needs the CLI.
- Backlogged: an explicit `Custom` Ethernet preset is reseeded to a named one when its pins match, because the seed cannot tell "not chosen" from "chose Custom". An attempted fix broke the legacy upgrade path and was reverted; the entry names the seam.
- Lessons: a control test must fail against the bug it claims to catch. Two tests passed against this commit's own cache bug before a third fired.

**Reviews**
- 🐇 The stride cache could not answer the test it was asked → replaced with the coordinates, pinned by a test that fails against the old cache.
- 🐇 `keptStep_` missing from the heap accounting → counted.
- 🐇 An unmatched preview point defaulted to color 0 → draws dark, and the index is bounds-checked.
- 🐇 Hard-wrap findings keyed repo-relative → docs-relative, matching the catalog.
- 🐇 A stale "click to toggle" tooltip → describes the hold gesture.
- 🐇 Four spellings and punctuation findings → fixed.
- 🐇 The `ethBoard` Custom conflation → verified real, fix reverted as worse than the defect, backlogged with its seam named.
- 🐇 A repo-health delta → generated output, rewritten by collect_kpi rather than edited.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 CLAUDE.md                               |  2 +
 docs/how-to/building.md                 |  2 +-
 docs/moonmodules/core/services.md       |  2 +-
 docs/moonmodules/core/ui.md             |  2 +-
 docs/moonmodules/light/layouts.md       |  2 +-
 docs/moonmodules/light/modifiers.md     |  2 +-
 docs/reference/MIGRATING.md             |  2 +-
 docs/reference/metrics/docgen.md        | 12 ++--
 docs/reference/metrics/repo-health.json | 42 +++++++-------
 docs/reference/metrics/repo-health.md   | 36 ++++++------
 docs/work/future/backlog-core.md        | 11 +++-
 docs/work/past/lessons.md               |  8 +++
 moondeck/check/check_docgen.py          |  4 +-
 mooninstaller/install-orchestrator.js   | 76 +++++++++----------------
 mooninstaller/install.js                | 30 ++--------
 src/core/system/NetworkModule.h         |  2 +-
 src/light/drivers/PreviewDriver.h       | 43 ++++++++------
 src/ui/app.js                           |  6 +-
 src/ui/preview3d.js                     | 12 +++-
 test/js/installer-s31-webflash.test.mjs | 47 ++++++++-------
 test/unit/light/unit_PreviewDriver.cpp  | 32 +++++++++--
 21 files changed, 202 insertions(+), 173 deletions(-)

diff --git a/CLAUDE.md b/CLAUDE.md
index 2779b40c..94407d9a 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -210,6 +210,8 @@ Each name is a script under `moondeck/`, run through `uv run`; the command and w
 
 **`check_docgen` is a ratchet.** Errors are resolved before a commit, and warnings may only fall: the committed `docs/reference/metrics/docgen.md` is the number to beat, on the total and on every rule. It fails a run that raises either. Per rule as well as per total, because a total hides one rule paying for another, and because the cheapest way to satisfy a width rule is to split a line, which raises the block count and fixes nothing. A rule whose own limit changed is the one case to say so in the commit.
 
+**A file with warnings is left better than it was found.** Holding the line is the floor, not the goal: the report is meant to shrink, and it only does so if each change spends a little effort on the warnings in the files it already touches. Reasonable effort, in the spirit of principle 5: the ones a reader would agree with, not a rewrite of every comment in the file. What resists is left with its count unchanged rather than forced, since a comment split to satisfy a width rule is the move the per-rule ratchet exists to refuse. Files the change never opened are a sweep of their own.
+
 Three checks earn their place for a reason worth knowing. **Repo health** is the only place the creeping numbers are visible: flash and DRAM per target, binary size, the tick matrix, line counts, complexity warnings. Its diff belongs in the commit and its deltas in the commit message. It runs when the code changes rather than on every commit, because its timings drift with the host: on a docs-only diff it records a regression that nothing in the diff caused. **The no-backend build** catches a helper left unused outside its guard, fatal under GCC while clang stays silent. **ESP32 firmware fresh** compares the binary against every source in a tenth of a second and catches the edit that was never compiled; compile for real after an sdkconfig or toolchain change. The [provisioning path](moondeck/MoonDeck.md#improv_smoke_test) is the five files MoonDeck names.
 
 
diff --git a/docs/how-to/building.md b/docs/how-to/building.md
index e131ffb7..968d9f08 100644
--- a/docs/how-to/building.md
+++ b/docs/how-to/building.md
@@ -177,7 +177,7 @@ and `flash_esp32.py` writes the corrected layout in one pass: app in the big `ot
 (cd ~/esp/esp-idf && ./install.sh esp32s31)   # one-time, adds the S31 RISC-V toolchain
 ```
 
-Flash the S31 over USB with the CLI (`flash_esp32.py --firmware esp32s31 --port <port>`), **not** the web installer: the browser flasher (`esptool-js`) has no S31 chip definition, so a browser flash fails, but the CLI's `esptool.py` supports it. The web installer surfaces the same guidance if you try. (Status + the condition to enable web flashing: [backlog](../work/future/index.md).)
+Flash the S31 from the web installer or with the CLI (`flash_esp32.py --firmware esp32s31 --port <port>`), whichever suits. Browser flashing works from esptool-js 0.7.0, which identifies the chip by its id rather than by a ROM magic value the S31 shares with the classic ESP32.
 
 On Windows, the `--port` argument is a `COM*` name (e.g. `COM3`) instead of `/dev/tty.usbserial-XXXX`. MoonDeck's port picker enumerates `COM*` automatically.
 
diff --git a/docs/moonmodules/core/services.md b/docs/moonmodules/core/services.md
index 11565f0e..46261672 100644
--- a/docs/moonmodules/core/services.md
+++ b/docs/moonmodules/core/services.md
@@ -162,7 +162,7 @@ Sending and receiving both use the **multicast address 239.0.0.1**, which is wha
 **Port 11988 is the WLED contract**, and `syncPort` defaults to it. The port is configurable for
 projectMM peers that want a private stream, but a custom port is no longer WLED-compatible: the endpoint WLED speaks is 239.0.0.1:11988 specifically.
 
-Multicast is also the better neighbour, with a caveat worth knowing: a switch or access point that does **IGMP snooping** forwards the group only to the ports that joined it, so the other hosts never see the traffic at all. Without snooping the switch floods it exactly like broadcast, and on WiFi it goes out at the lowest basic rate to every station. So multicast can reduce how many hosts have to process ~40 packets a second, but it does not guarantee it. See [multicast and IGMP snooping](../../explanation/architecture/moonlight.md#multicast-and-igmp-snooping).
+Multicast is also the better neighbor, with a caveat worth knowing: a switch or access point that does **IGMP snooping** forwards the group only to the ports that joined it, so the other hosts never see the traffic at all. Without snooping the switch floods it exactly like broadcast, and on WiFi it goes out at the lowest basic rate to every station. So multicast can reduce how many hosts have to process ~40 packets a second, but it does not guarantee it. See [multicast and IGMP snooping](../../explanation/architecture/moonlight.md#multicast-and-igmp-snooping).
 
 The 44-byte v2 packet is byte-compatible with WLED, with one field that is not yet equivalent:
 
diff --git a/docs/moonmodules/core/ui.md b/docs/moonmodules/core/ui.md
index 40f44cf7..cbafa7fb 100644
--- a/docs/moonmodules/core/ui.md
+++ b/docs/moonmodules/core/ui.md
@@ -163,7 +163,7 @@ One picker serves **add** (`+ add child`) and **replace** (the ✎ button), rend
 
 ## Domain preview channel
 
-The UI dedicates a binary slot on its WebSocket, separate from the JSON state updates, for a domain-specific preview frame; the engine pushes one per render and the UI hands it to a domain renderer. Generic shape: `[type-byte] [domain header] [payload]`, the first byte selects the renderer, everything after is the domain's choice; the UI ignores types it doesn't recognise.
+The UI dedicates a binary slot on its WebSocket, separate from the JSON state updates, for a domain-specific preview frame; the engine pushes one per render and the UI hands it to a domain renderer. Generic shape: `[type-byte] [domain header] [payload]`, the first byte selects the renderer, everything after is the domain's choice; the UI ignores types it doesn't recognize.
 
 - The canvas is sticky just below the status bar, scroll-shrinks 0→50 % over 300 px.
 - `width: 100%` + `aspect-ratio: 1 / 1` derives the height from the column width; `max-height: 50vh`
diff --git a/docs/moonmodules/light/layouts.md b/docs/moonmodules/light/layouts.md
index 10dc3697..2c3edf84 100644
--- a/docs/moonmodules/light/layouts.md
+++ b/docs/moonmodules/light/layouts.md
@@ -228,7 +228,7 @@ Detail: [technical](moxygen/MoonLiveLayout.md) · [how the count is known](#moon
 
 A dense 3D grid, row-major (x fastest, then y, then z); every position maps to a light.
 
-- `width` / `height` / `depth` — lights per axis (to 3840, 2160 and 512).
+- `width` / `height` / `depth`: lights per axis (to 3840, 2160 and 512).
 - `serpentine` — every other row runs in reverse, matching a snaked strip.
 
 Origin: MoonLight · via [MoonLight](https://github.com/MoonModules/MoonLight/blob/main/src/MoonLight/Nodes/Layouts/L_MoonLight.h)
diff --git a/docs/moonmodules/light/modifiers.md b/docs/moonmodules/light/modifiers.md
index f59a6331..0e24b5ed 100644
--- a/docs/moonmodules/light/modifiers.md
+++ b/docs/moonmodules/light/modifiers.md
@@ -1,6 +1,6 @@
 # Modifiers
 
-Every modifier, one block each: its preview, what it does, and what each control means — together. A modifier sits between an [effect](effects.md) and the output: it reshapes *where* pixels land (or masks them) without changing the effect's drawing. Modifiers compose — a [Layer](moxygen/Layer.md) folds its whole modifier stack each rebuild; a *dynamic* modifier (one that overrides `modifyLive`) also runs a per-frame pass, and an *animated* one keeps a static fold but rebuilds it on its own clock. See [ModifierBase](moxygen/ModifierBase.md) for the static-vs-dynamic split. Each block's emoji are its `tags()` (see the [tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend)); **Kind** is static (baked into the mapping at rebuild) or dynamic (per-frame remap). Modifiers are grouped into sections, and each block carries that modifier's preview, behavior, and control descriptions together. (For how this page maps to the source/asset folders, see the [folder-structure decision](../../contributing/documentation-standards.md#module-pages).)
+Every modifier, one block each: its preview, what it does, and what each control means, together. A modifier sits between an [effect](effects.md) and the output: it reshapes *where* pixels land (or masks them) without changing the effect's drawing. Modifiers compose: a [Layer](moxygen/Layer.md) folds its whole modifier stack each rebuild; a *dynamic* modifier (one that overrides `modifyLive`) also runs a per-frame pass, and an *animated* one keeps a static fold but rebuilds it on its own clock. See [ModifierBase](moxygen/ModifierBase.md) for the static-vs-dynamic split. Each block's emoji are its `tags()` (see the [tag emoji legend](../../explanation/architecture/index.md#tag-emoji-legend)); **Kind** is static (baked into the mapping at rebuild) or dynamic (per-frame remap). Modifiers are grouped into sections, and each block carries that modifier's preview, behavior, and control descriptions together. (For how this page maps to the source/asset folders, see the [folder-structure decision](../../contributing/documentation-standards.md#module-pages).)
 
 A modifier folds coordinates rather than drawing, so it reaches for very little of the shared [power function](power-functions.md) toolbox — that page states the split and lists which modifiers use what.
 
diff --git a/docs/reference/MIGRATING.md b/docs/reference/MIGRATING.md
index ebc21645..86fc33be 100644
--- a/docs/reference/MIGRATING.md
+++ b/docs/reference/MIGRATING.md
@@ -251,7 +251,7 @@ The `peripheral` dropdown no longer says `i80` / `MoonI80`. "i80" is the Intel 8
 | Old | New (classic ESP32) | New (S3 / P4 / S31) |
 |---|---|---|
 | `i80` | `I2S-IDF` | `LCD-IDF` |
-| `MoonI80` |, (not available) | `LCD-MM` |
+| `MoonI80` | (not available) | `LCD-MM` |
 | `Parlio` | — | `Parlio` (unchanged, it *is* the peripheral's name) |
 
 `-IDF` = driven through ESP-IDF's `esp_lcd`; `-MM` = driven by our own GDMA layer below it, which is what buys the streaming ring and the 74HCT595 pin expander.
diff --git a/docs/reference/metrics/docgen.md b/docs/reference/metrics/docgen.md
index 4b638757..a576694a 100644
--- a/docs/reference/metrics/docgen.md
+++ b/docs/reference/metrics/docgen.md
@@ -4,7 +4,7 @@ Generated by [`moondeck/check/check_docgen.py`](../../../moondeck/check/check_do
 
 Every place the generated documentation breaks the shape [the standards](../../contributing/documentation-standards.md#the-card) define. Current state only: the trend is this file's git history. The list only shrinks.
 
-**0 error(s)** and **3135 warning(s)** across 217 page(s).
+**0 error(s)** and **3131 warning(s)** across 217 page(s).
 
 An error is in a file that generates a documentation page, a header or a catalog page, so the finding is a defect in what gets published and it fails the gate. A warning is in an implementation file, which publishes nothing: its comments are a note to the next reader, worth fixing without being worth stopping a commit for. Both are counted here, because a warning nobody sees is a warning nobody fixes.
 
@@ -14,7 +14,7 @@ The split is temporary. It stages the sweep rather than ranking the two kinds of
 
 | Rule | Errors | Warnings |
 |---|---:|---:|
-| over-wide comment lines | 0 | 2229 |
+| over-wide comment lines | 0 | 2225 |
 | multi-line comment blocks | 0 | 604 |
 | over-long sentences | 0 | 108 |
 | over-wide doc lines | 0 | 102 |
@@ -26,7 +26,7 @@ The unit a sweep runs in: one summary page and the headers it owns, as the [hier
 
 | Summary page | Errors | Warnings |
 |---|---:|---:|
-| `(tests, no card)` | 0 | 2207 |
+| `(tests, no card)` | 0 | 2203 |
 | `platform/index.md` | 0 | 518 |
 | `core/system.md` | 0 | 410 |
 
@@ -36,7 +36,7 @@ Per area, since that is the unit a sweep runs in: the files ranked, then the rul
 
 ### (tests, no card)
 
-**0 error(s)** and **2207 warning(s)** across 184 file(s).
+**0 error(s)** and **2203 warning(s)** across 184 file(s).
 
 | Findings | File | |
 |---:|---|---|
@@ -50,9 +50,9 @@ Per area, since that is the unit a sweep runs in: the files ranked, then the rul
 | 44 | `test/unit/core/unit_moonlive_codegen_xtensa.cpp` | warning |
 | 42 | `test/unit/light/unit_ParallelSlots.cpp` | warning |
 | 41 | `test/scenario_runner.cpp` | warning |
-| 1-41 each | *174 more warning files, 1541 findings* | |
+| 1-41 each | *174 more warning files, 1537 findings* | |
 
-By rule, warnings: 1672 over-wide comment lines, 239 multi-line comment blocks, 106 over-long sentences, 101 over-wide doc lines, 89 member deep dives.
+By rule, warnings: 1668 over-wide comment lines, 239 multi-line comment blocks, 106 over-long sentences, 101 over-wide doc lines, 89 member deep dives.
 
 ### platform/index.md
 
diff --git a/docs/reference/metrics/repo-health.json b/docs/reference/metrics/repo-health.json
index 36bd0673..d27ed1b0 100644
--- a/docs/reference/metrics/repo-health.json
+++ b/docs/reference/metrics/repo-health.json
@@ -1,13 +1,13 @@
 {
-  "commit": "f1be21c4",
+  "commit": "8f50ae79",
   "flash": {
     "esp32s3-n16r8": 2134032,
-    "desktop": 1970664,
-    "esp32": 2084672,
-    "esp32p4rev1-eth": 2034288,
+    "desktop": 1970712,
+    "esp32": 2084912,
+    "esp32p4rev1-eth": 2034592,
     "esp32p4rev1-eth-wifi": 2284640,
     "esp32s3-n8r8": 2087168,
-    "esp32s31": 2427440,
+    "esp32s31": 2427632,
     "esp32-16mb": 2060368,
     "esp32-eth": 1681040,
     "esp32-wrover": 1843760,
@@ -31,8 +31,8 @@
   },
   "perf": {
     "desktop": {
-      "tick_us": 124,
-      "fps": 8064,
+      "tick_us": 121,
+      "fps": 8264,
       "scenario_p50": {
         "Layer_base_pipeline": {
           "p50": 70,
@@ -699,11 +699,11 @@
   },
   "loc": {
     "core": 22010,
-    "light": 29952,
+    "light": 29961,
     "platform": 16346,
-    "ui": 11310,
-    "test": 56377,
-    "moondeck": 27235
+    "ui": 11318,
+    "test": 56401,
+    "moondeck": 27237
   },
   "comments": {
     "core": {
@@ -711,7 +711,7 @@
       "ratio": 0.28
     },
     "light": {
-      "lines": 7269,
+      "lines": 7270,
       "ratio": 0.272
     },
     "platform": {
@@ -719,32 +719,32 @@
       "ratio": 0.238
     },
     "ui": {
-      "lines": 3371,
+      "lines": 3376,
       "ratio": 0.315
     },
     "test": {
-      "lines": 6360,
+      "lines": 6362,
       "ratio": 0.131
     },
     "moondeck": {
-      "lines": 4570,
+      "lines": 4572,
       "ratio": 0.191
     }
   },
   "tests": {
-    "cases": 2070,
+    "cases": 2071,
     "scenarios": 27
   },
   "docs": {
     "md_files": 138,
-    "md_lines": 28265,
+    "md_lines": 28282,
     "plans_files": 31,
-    "backlog_lines": 3025,
-    "lessons_lines": 518,
-    "claude_md_lines": 279
+    "backlog_lines": 3034,
+    "lessons_lines": 526,
+    "claude_md_lines": 281
   },
   "complexity": {
-    "functions": 3630,
+    "functions": 3631,
     "over_threshold": 267,
     "worst_ccn": 128
   }
diff --git a/docs/reference/metrics/repo-health.md b/docs/reference/metrics/repo-health.md
index 5178a840..b9c21c4a 100644
--- a/docs/reference/metrics/repo-health.md
+++ b/docs/reference/metrics/repo-health.md
@@ -1,6 +1,6 @@
 # Repo health
 
-Measured at `f1be21c4`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
+Measured at `8f50ae79`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
 
 Current state only; the trend is this file's git history (`git log -p docs/reference/metrics/repo-health.md`). Nothing here fails a build: the numbers make growth visible, the judgment stays human.
 
@@ -8,19 +8,19 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Flash | Capacity | Used | Built |
 |---|---:|---:|---:|:--:|
-| desktop | 1,924 KB (+96 B) ⚠ | - | - | yes |
-| esp32 | 2,036 KB (+4 KB) ⚠ | 2,496 KB | 82% | yes |
+| desktop | 1,925 KB (+48 B) ⚠ | - | - | yes |
+| esp32 | 2,036 KB (+240 B) ⚠ | 2,496 KB | 82% | yes |
 | esp32-16mb | 2,012 KB | 4,096 KB | 49% | **STALE 12d** |
 | esp32-eth | 1,642 KB | 2,496 KB | 66% | **STALE 10d** |
 | esp32-pico | 2,058 KB | 3,072 KB | 67% | **STALE 12d** |
 | esp32-wrover | 1,801 KB | - | - | carried (age?) |
-| esp32p4rev1-eth | 1,987 KB (+4 KB) ⚠ | 4,096 KB | 49% | yes |
+| esp32p4rev1-eth | 1,987 KB (+304 B) ⚠ | 4,096 KB | 49% | yes |
 | esp32p4rev1-eth-wifi | 2,231 KB | 4,096 KB | 54% | **STALE 13d** |
 | esp32p4rev3-eth | 1,605 KB | - | - | carried (age?) |
 | esp32s3-n16r8 | 2,084 KB | 4,096 KB | 51% | yes |
 | esp32s3-n8r8 | 2,038 KB | 3,072 KB | 66% | **STALE 13d** |
 | esp32s3-zero | 1,977 KB | 2,496 KB | 79% | **STALE 13d** |
-| esp32s31 | 2,371 KB (+77 KB) ⚠ | 4,096 KB | 58% | yes |
+| esp32s31 | 2,371 KB (+192 B) ⚠ | 4,096 KB | 58% | yes |
 | qemu | 1,351 KB | - | - | carried (age?) |
 
 `Built: yes` was measured this run. `carried (age?)` was not rebuilt either and predates this record, so its age is unknown: it dates itself on the next build. `carried Nd` was NOT rebuilt and its number is N days old, so an absent delta says nothing about the change. **STALE** marks a carry older than 7 days: the number has gone unchecked long enough that growth will surface later as one jump, blamed on whichever commit happens to rebuild that target. `Used` is against the app slot in the firmware's own partition table.
@@ -29,7 +29,7 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Tick | FPS |
 |---|---:|---:|
-| desktop | 124 µs (−11 µs) ✓ | 8,064 (+657) ✓ |
+| desktop | 121 µs (−3 µs) ✓ | 8,264 (+200) ✓ |
 | esp32 | 8,354 µs | 119 |
 
 ### Scenario tick by target (p50 of each sample window)
@@ -81,26 +81,26 @@ These build a bare pipeline with no optional modules, so a change here is a chan
 
 | Area | Lines | Comments | Comment share |
 |---|---:|---:|---:|
-| core | 22,010 (+16) ⚠ | 5,579 | 28.0 % |
-| light | 29,952 (+49) ⚠ | 7,269 | 27.2 % (−0.1 %) ✓ |
+| core | 22,010 | 5,579 | 28.0 % |
+| light | 29,961 (+9) ⚠ | 7,270 | 27.2 % |
 | platform | 16,346 | 3,481 | 23.8 % |
-| ui | 11,310 (+87) ⚠ | 3,371 | 31.5 % |
-| test | 56,377 (+112) ⚠ | 6,360 | 13.1 % |
-| moondeck | 27,235 (+122) ⚠ | 4,570 | 19.1 % |
+| ui | 11,318 (+8) ⚠ | 3,376 | 31.5 % |
+| test | 56,401 (+24) ⚠ | 6,362 | 13.1 % |
+| moondeck | 27,237 (+2) ⚠ | 4,572 | 19.1 % |
 
 ## Tests
 
 | Kind | Count |
 |---|---:|
-| unit cases | 2,070 (+3) ✓ |
+| unit cases | 2,071 (+1) ✓ |
 | scenarios | 27 |
 
 ## Complexity
 
 | Metric | Value |
 |---|---:|
-| functions | 3,630 (+3) ✓ |
-| over threshold | 267 (+2) ⚠ |
+| functions | 3,631 (+1) ✓ |
+| over threshold | 267 |
 | worst CCN | 128 |
 
 ## Documentation
@@ -108,9 +108,9 @@ These build a bare pipeline with no optional modules, so a change here is a chan
 | Metric | Value |
 |---|---:|
 | markdown files | 138 |
-| markdown lines | 28,265 (−721) ✓ |
+| markdown lines | 28,282 (+17) ⚠ |
 | plan files | 31 |
-| backlog lines | 3,025 |
-| lessons lines | 518 |
-| CLAUDE.md lines | 279 |
+| backlog lines | 3,034 (+9) ⚠ |
+| lessons lines | 526 (+8) ⚠ |
+| CLAUDE.md lines | 281 (+2) ⚠ |
 
diff --git a/docs/work/future/backlog-core.md b/docs/work/future/backlog-core.md
index 7fc0f6b8..fe7386a0 100644
--- a/docs/work/future/backlog-core.md
+++ b/docs/work/future/backlog-core.md
@@ -130,7 +130,6 @@ declared rather than for a buffer to fill, and to time out on stall rather than
 1.0 ships ESP32 firmware (4 variants) + macOS arm64 + Windows x64. Still to add:
 
 - **ESP32-P4** firmware variant — **`esp32p4rev1-eth` (Ethernet-only) shipped**: in `build_esp32.py`'s `FIRMWARES`, the `deviceModels.json` catalog (Waveshare P4-NANO), and CI builds + publishes it to the web installer + releases. **`esp32p4rev1-eth-wifi` now ships too** (2026-08-19): it boots and associates on IDF v6.1-rc1, so it is out of the experimental set in the installer and carries a normal description. Its open defect is throughput, not shipping — see § ESP32-P4 round 3, open issue 0.
-- **ESP32-S31 web-flash (waiting on esptool-js)** — the `esp32s31` firmware ships (build, catalog, CI matrix, web installer listing), and CLI flashing works (`flash_esp32.py` → esptool.py, which has S31 support since v5.2.0). **Browser flashing does not**: the web installer's `esptool-js` (pinned 0.5.7) has no S31 chip class. Worse than a missing entry — the S31's ROM magic (`15736195`) *collides* with the classic ESP32's; esptool.py disambiguates with secondary register detection (S31 `USES_MAGIC_VALUE=False`), but esptool-js has only the magic table, so it would mis-identify the RISC-V S31 as a classic Xtensa ESP32 and flash the wrong stub/params. `install.js`'s `WEB_FLASH_UNSUPPORTED_CHIPS` guard catches an S31 connect-flash failure and points the user at the CLI. **No upstream timeline**: as of 2026-06 the esptool-js repo has zero S31 issues/PRs/commits and its last release was 2026-03 (it lags esptool.py on new chips by months). **Removal trigger**: when esptool-js ships S31 support *with* the secondary detection (not just a magic-table entry — re-check the chip-detect switch, not the version number), bump the esptool-js pin in `install-orchestrator.js` and drop `ESP32-S31` from `WEB_FLASH_UNSUPPORTED_CHIPS`.
 - **ESP32-P4 v3.x silicon variant (backlog)** — `esp32p4rev1-eth` is built for pre-v3 P4 (`CONFIG_ESP32P4_SELECTS_REV_LESS_V3` + `REV_MIN_0`), because the v6.1 IDF default (v3.1) refused to boot on the bench/field v1.x P4 and rev <3.0 vs >=3.0 are "huge hardware difference" (one binary can't cover both). **DONE (2026-08-19): `esp32p4rev3-eth` and `esp32p4rev3-eth-wifi` ship** (`REV_MIN_300`, which covers v3.0-v3.99), reusing the rev1 board fragment so the partition table and EMAC config stay in one place. **Still open: neither has ever been booted** — both bench boards are v1.3 engineering samples, so the rev3 images are flagged experimental in the installer and need a v3 board to verify. Espressif does not recommend v0.x/v1.x for new designs, so a board bought today is v3.x and needs these.
 - **Teensy 4.1** — toolchain-file build, `.hex` for Teensy Loader.
 - **Raspberry Pi** — ARM64, cross-built or native.
@@ -832,6 +831,16 @@ Run a bare-leaving scenario before a tree-assuming one and the latter fails pre-
 
 Fix options: (a) make every live mutate scenario clear+rebuild its own canvas (consistent with the newer ones) so order never matters; or (b) have the live runner reboot / restore the canonical tree between scenarios. (a) is the cleaner long-term shape. Until then, the in-process suite is the gate; live full-suite runs need a clean boot per scenario, or run scenarios individually.
 
+## An explicit `Custom` Ethernet preset is reseeded to a named one
+
+`ethPresetIsUnset()` returns true for any editable row, and `Custom` is the only editable row, so it cannot tell "nothing has chosen a preset" from "the user chose Custom". A board saved as `Custom` whose pins happen to match a named preset is therefore reseeded to that preset on the next rebuild, and its pin rows are hidden: a hand-wired board silently adopts a board map it was never given.
+
+The seam is what makes this awkward. `defineControls` runs before the config is restored, so the seed cannot simply latch a flag on the first render: that is the legacy path the seed exists for, a board whose file predates `ethBoard` and carries only pins. Distinguishing the two needs to know whether the restore supplied an `ethBoard` value at all, which is not visible where the seed runs. An attempt that latched on the first render broke the legacy path and was reverted.
+
+**What it costs today**: nothing on a board the catalog configures, since those name their preset. It bites a user who deliberately picks `Custom` on a board whose pins match a preset, which is the hand-wired case `Custom` exists for.
+
+**Where to look**: `FilesystemModule::applyNode` overlays values, calls `rebuildControls`, then overlays again. A control that wants to know whether IT was restored needs a signal from that pass, which no control has today. That is the general shape worth solving rather than a special case for one select.
+
 ## Ethernet on the QuinLED Dig-Next-2 (W5500 over SPI)
 
 The DN2 is in the catalog with no Ethernet block, so it comes up WiFi-only. Adding it is a `deviceModels.json` entry and nothing else: the preset work landed the seam, and W5500 already applies live.
diff --git a/docs/work/past/lessons.md b/docs/work/past/lessons.md
index cb08cdbb..75878397 100644
--- a/docs/work/past/lessons.md
+++ b/docs/work/past/lessons.md
@@ -516,3 +516,11 @@ frame when the previous one has not drained, report the skip to the receiver (it
 congestion signal), and close only on a real error or FIN. The receiver steers quality from the
 drop reports. Every give-up budget that remains must bound *lack of progress*, never elapsed
 total, or slow-but-healthy transfers get truncated.
+
+## A control test must fail against the bug it claims to catch
+
+A reviewer flagged that caching one "coarsest divisor" per light could not answer a stride test, and the finding was right. The reasoning offered with it was not: it named (70,70,0) at stride 2, but 70's gcd is 70 and the divisor loop stops at 64, so that coordinate is unaffected. The real condition is narrower, a gcd ABOVE the loop's cap, which no coordinate under 65 can reach.
+
+Two tests were written and both passed against the unfixed code before one finally fired. The first used a dense grid, which walks the lattice closed-form and never reads the cache. The second used a sphere whose coordinates all sat under the cap, so both predicates agreed on every light it contained. Only a radius-64 shell, which reaches coordinates whose axes share a factor above 64, separated them.
+
+**The rule: reproduce the failure before fixing it, and prove the test fails against the old code.** A finding's headline can be sound while its example is wrong, so the example is the first thing to verify, not the claim. And a test that passes against the bug is not a weaker test, it is not a test of that bug at all. The control run, old code plus new test, is what tells the difference, and it costs one build.
diff --git a/moondeck/check/check_docgen.py b/moondeck/check/check_docgen.py
index 1a72513f..efca65b5 100755
--- a/moondeck/check/check_docgen.py
+++ b/moondeck/check/check_docgen.py
@@ -1417,7 +1417,9 @@ def _violations():
         # moxygen writes generated pages; docs/work/ is the exemption the standards name.
         if "moxygen" in parts or parts[0] == "work":
             continue
-        out.extend(_md_hard_wraps(rel.relative_to(ROOT).as_posix(), rel.read_text()))
+        # DOCS-relative, the key catalog findings use: a repo-relative one listed the same
+        # page twice in the report, once per spelling.
+        out.extend(_md_hard_wraps(rel.relative_to(ROOT / "docs").as_posix(), rel.read_text()))
 
     for rel in _headers():
         out.extend(_header_rules(str(rel), (ROOT / rel).read_text()))
diff --git a/mooninstaller/install-orchestrator.js b/mooninstaller/install-orchestrator.js
index e28f146f..6a9b628d 100644
--- a/mooninstaller/install-orchestrator.js
+++ b/mooninstaller/install-orchestrator.js
@@ -31,33 +31,29 @@
 // flash failure can be tied to the exact flasher build. The version constants
 // below MUST match the import URLs (a check script could pin this later).
 //
-// esptool-js is pinned 0.5.7 — the version ESP Web Tools (the flasher ESPHome
-// and WLED embed) ships. 0.6.0 (the newest, tagged 2026-03-26) has a DETERMINISTIC
-// compressed-flash bug: a P4 web-flash aborts at a FIXED block — "Failed to write
-// compressed data to flash after seq NN failed with status 201,0" — where 0.4.7,
-// 0.5.7, and the CLI (esptool.py) all flash the same P4 cleanly. Failing at a fixed
-// seq (not a random one) rules out a transient USB hiccup; it's a real 0.6.0
-// regression in the deflate write path (cf. upstream esptool-js#233/#245). 0.5.x
-// also moved hardReset off ESPLoader into a reset-strategy class — handled by
-// hardResetChip() below (transport DTR/RTS), version-agnostic, so 0.6.x would
-// reboot fine IF its flash worked. Re-test the flash + reset path on any bump;
-// 0.6.x is only viable once that deflate regression is fixed.
-//   Re-verified on the bench 2026-07-27: 0.6.0 is still the newest tag (no 0.6.1+),
-//   and a real P4 web-flash STILL aborts — "seq 50 failed with status 201,0". So the
-//   regression persists in 0.6.0-as-tagged; keep 0.5.7. (0.6.0 also brings no ESP32-S31
-//   support — misdetection is tracked upstream in esptool-js#248 — so the bump has no
-//   upside for us either.) Pinned 2026-06-28, re-verified 2026-07-27.
-//   Re-checked 2026-08-19: **0.6.1 shipped (2026-08-06) and its notes name the fix we
-//   are waiting on** — upstream #245 "Add retries to FLASH_DATA and FLASH_DEFL_DATA",
-//   plus #244 (uncompressed data in writeFlash) and #249 (connection reliability). That
-//   is the deflate write path this pin exists to avoid, so 0.6.1 is the first bump worth
-//   a real P4 bench flash. NOT yet tested here — the pin stays 0.5.7 until a P4
-//   web-flash completes on 0.6.1. Still no ESP32-S31 support in 0.6.1 (that is separate,
-//   see WEB_FLASH_UNSUPPORTED_CHIPS in install.js), so the S31 CLI path is unaffected
-//   either way.
-export const ESPTOOL_JS_VERSION = "0.5.7";
+// esptool-js is pinned 0.7.0 (2026-09-21), the first release that flashes an ESP32-S31 from a
+// browser. Detection now asks the chip its id (GET_SECURITY_INFO, upstream #197) before falling
+// back to the magic register, which is what makes the S31 safe: its ROM magic COLLIDES with the
+// classic ESP32's, so a magic-only table mis-identified the RISC-V part as an Xtensa one and would
+// have flashed the wrong stub. `esp32s31.ts` sets USES_MAGIC_VALUE=false with IMAGE_CHIP_ID 32, the
+// same way esptool.py disambiguates, and ships its own stub_flasher/esp32s31.json.
+//
+// 0.6.0 was skipped for a DETERMINISTIC compressed-flash bug: a P4 web-flash aborted at a fixed
+// block ("failed with status 201,0") where 0.5.7 and the CLI flashed the same board cleanly.
+// 0.6.1 named the fix (#245, retries on FLASH_DATA/FLASH_DEFL_DATA) and 0.7.0 adds #268, which
+// calls powerOnFlash() from postConnect(): on ECO6/ECO7 P4 silicon the flash is powered off by
+// default, so the flash-ID read returned garbage and the first flash command hung the stub.
+//
+// Two breaking changes came with it, both handled here: writeFlash THROWS unless each part's
+// `data` is a Uint8Array (a binary string silently corrupted the image, #266), and
+// detectFlashSize() returns undefined rather than defaulting to "4MB" — unused here, since the
+// flash call passes flashSize "keep".
+//
+// Re-test the flash + reset path on any bump; hardResetChip() below drives DTR/RTS through the
+// transport, so it stays version-agnostic.
+export const ESPTOOL_JS_VERSION = "0.7.0";
 export const IMPROV_SDK_VERSION = "2.5.0";
-import { ESPLoader, Transport } from "https://unpkg.com/esptool-js@0.5.7/bundle.js?module";
+import { ESPLoader, Transport } from "https://unpkg.com/esptool-js@0.7.0/bundle.js?module";
 import { ImprovSerial } from "https://unpkg.com/improv-wifi-serial-sdk@2.5.0/dist/serial.js?module";
 
 // ---------------------------------------------------------------------------
@@ -116,24 +112,6 @@ async function fetchManifest(manifestUrl) {
     };
 }
 
-// Convert an ArrayBuffer to a "binary string" — one JS character per byte,
-// codes 0x00-0xFF. esptool-js's writeFlash expects each fileArray entry's
-// `data` in this shape (it iterates via .charCodeAt()). Chunked at 16 KB
-// because `String.fromCharCode(...big_array)` blows the call stack on
-// large inputs (a 1.2 MB app image would otherwise spread ~1.2M arguments).
-function bufferToBinaryString(buffer) {
-    const bytes = new Uint8Array(buffer);
-    const CHUNK = 16384;
-    let out = "";
-    for (let i = 0; i < bytes.length; i += CHUNK) {
-        out += String.fromCharCode.apply(
-            null,
-            bytes.subarray(i, Math.min(i + CHUNK, bytes.length))
-        );
-    }
-    return out;
-}
-
 // ---------------------------------------------------------------------------
 // Improv RPC payload encoders (frame building is in improv-frame.js)
 // ---------------------------------------------------------------------------
@@ -747,13 +725,11 @@ export const installer = {
                     throw new Error(`part fetch HTTP ${res.status}: ${part.url}`);
                 }
                 const buf = await res.arrayBuffer();
-                // esptool-js 0.4.7's writeFlash expects each part's `data`
-                // as a binary string (one char per byte) — it iterates with
-                // .charCodeAt() inside. Uint8Array fails with "charCodeAt is
-                // not a function". Convert in 16 KB chunks to avoid blowing
-                // the call-stack on a 1.2 MB app image.
+                // esptool-js 0.7.0 THROWS unless `data` is a Uint8Array: a binary string used to
+                // be the required shape (it iterated with .charCodeAt()), and passing one now
+                // corrupts the image instead of flashing it (upstream #266).
                 fileArray.push({
-                    data: bufferToBinaryString(buf),
+                    data: new Uint8Array(buf),
                     address: part.offset,
                 });
             }
diff --git a/mooninstaller/install.js b/mooninstaller/install.js
index 9fbde554..7a8aa2f0 100644
--- a/mooninstaller/install.js
+++ b/mooninstaller/install.js
@@ -593,30 +593,12 @@ document.addEventListener('DOMContentLoaded', () => {
       myDevices.addProvisionedDevice(url, defaultsApplied ? board : "");
     }
 
-    // esptool-js (the browser flasher) has no chip definition for these targets
-    // yet, so a browser flash can't begin — the CLI (esptool.py, which does know
-    // them) is the working path. Drop a chip from this set once esptool-js ships
-    // support for it (then also bump the esptool-js pin in install-orchestrator.js).
-    //
-    // The S31 is doubly unsafe in the browser, not just unsupported: its chip
-    // magic (0xF01D2E07 / 15736195) COLLIDES with the classic ESP32's. esptool.py
-    // disambiguates them with secondary register detection (it sets the S31's
-    // USES_MAGIC_VALUE=False); esptool-js 0.6.0 has only the magic table, where
-    // that value maps to ESP32ROM. So even if the S31's ROM-bootloader sync ever
-    // succeeded, esptool-js would mis-identify the RISC-V S31 as a classic Xtensa
-    // ESP32 and try to flash it with the wrong stub + flash params — corruption,
-    // not a lucky success. esptool-js needs the S31 secondary-detection logic
-    // before browser flashing is safe; a version bump alone is not enough.
-    //
-    // Re-verified 2026-08-19 against esptool-js main and the 0.6.1 release (2026-08-06):
-    // still NO S31. src/targets/ has no esp32s31.ts (esp32{,c2,c3,c5,c6,c61,h2,p4,s2,s3}.ts
-    // + esp8266/rom only), esploader.ts's magic2Chip has no ESP32S31ROM entry, and a repo
-    // code search for "S31" returns nothing. Support exists only as OPEN PR
-    // esptool-js#250, which adds ESP32S31ROM and — confirming the collision analysis above
-    // — identifies the chip via GET_SECURITY_INFO rather than the magic table; it is
-    // blocked on overlapping unreviewed work in #197. esp-web-tools depends on
-    // esptool-js ^0.6.0, so it inherits the same gap. Watch #250 landing in a 0.6.2/0.7.0.
-    const WEB_FLASH_UNSUPPORTED_CHIPS = new Set(["ESP32-S31"]);
+    // Every chip projectMM ships is browser-flashable as of esptool-js 0.7.0, which added the
+    // ESP32-S31 target and chip-id detection (GET_SECURITY_INFO) so the S31's magic collision with
+    // the classic ESP32 can no longer mis-identify it. The set stays because a NEW chip lands here
+    // before esptool-js knows it, and an empty set is the honest way to say "none, today".
+    // The CLI (esptool.py) remains the fallback the error path points at.
+    const WEB_FLASH_UNSUPPORTED_CHIPS = new Set();
 
     // Map a firmware key to its chip family ("esp32s31" → "ESP32-S31", "esp32s3-n16r8"
     // → "ESP32-S3", "esp32p4rev1-eth" → "ESP32-P4", "esp32*" → "ESP32") — the same prefix
diff --git a/src/core/system/NetworkModule.h b/src/core/system/NetworkModule.h
index 56370d3c..aae4575d 100644
--- a/src/core/system/NetworkModule.h
+++ b/src/core/system/NetworkModule.h
@@ -544,7 +544,7 @@ class NetworkModule : public MoonModule {
     const char* ethPresetOptions_[kEthPresetCount] = {};
     uint8_t ethPresetIndex_[kEthPresetCount] = {};
     uint8_t ethPresetCount_ = 0;
-    uint8_t ethPresetApplied_ = 0;   ///< the selection whose map is already written, so a rebuild is not a re-apply
+    uint8_t ethPresetApplied_ = 0;
 
 
     // A preset naming a PHY this build cannot drive would offer pins that reach nothing.
diff --git a/src/light/drivers/PreviewDriver.h b/src/light/drivers/PreviewDriver.h
index a57afa11..801d754e 100644
--- a/src/light/drivers/PreviewDriver.h
+++ b/src/light/drivers/PreviewDriver.h
@@ -73,6 +73,12 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
 
 
 
+    /// Does this kept light survive a frame at `fs`? The device's own rule, on its own coordinates.
+    bool keptAtStride(nrOfLightsType k, nrOfLightsType fs) const {
+        const size_t at = static_cast<size_t>(k) * 3;
+        return keptPos_[at] % fs == 0 && keptPos_[at + 1] % fs == 0 && keptPos_[at + 2] % fs == 0;
+    }
+
     /// The stride a color frame ships at: the link's pacing, never finer than the table's.
     nrOfLightsType frameStride() const {
         return downscale_ > previewStride_ ? downscale_ : previewStride_;
@@ -92,9 +98,9 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
             return static_cast<nrOfLightsType>(static_cast<uint32_t>(cx) * cy * cz);
         }
         // From the CACHE, never a placeLights walk: this runs per frame on a nonblocking tick.
-        if (!keptStep_ || keptCount_ != coordCount_) return coordCount_;
+        if (!keptPos_ || keptCount_ != coordCount_) return coordCount_;
         nrOfLightsType out = 0;
-        for (nrOfLightsType k = 0; k < keptCount_; k++) if (keptStep_[k] % fs == 0) out++;
+        for (nrOfLightsType k = 0; k < keptCount_; k++) if (keptAtStride(k, fs)) out++;
         return out;
     }
 
@@ -132,12 +138,12 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
             ensureStaging(maxPts * 3u);
             if (!denseGrid() && keptIdxCap_ < maxPts) {
                 auto* grown = static_cast<nrOfLightsType*>(platform::alloc(maxPts * sizeof(nrOfLightsType)));
-                auto* steps = static_cast<uint8_t*>(platform::alloc(maxPts));
+                auto* steps = static_cast<uint8_t*>(platform::alloc(maxPts * 3));
                 if (grown && steps) {
                     if (keptIdx_) platform::free(keptIdx_);
-                    if (keptStep_) platform::free(keptStep_);
+                    if (keptPos_) platform::free(keptPos_);
                     keptIdx_ = grown;
-                    keptStep_ = steps;
+                    keptPos_ = steps;
                     keptIdxCap_ = static_cast<nrOfLightsType>(maxPts);
                     publishHeapBytes();
                 } else {
@@ -273,12 +279,12 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
             // Sized to EXACTLY this count before the emit fills it, so the cache cannot truncate.
             if (keptIdxCap_ < coordCount_) {
                 auto* grown = static_cast<nrOfLightsType*>(platform::alloc(coordCount_ * sizeof(nrOfLightsType)));
-                auto* steps = static_cast<uint8_t*>(platform::alloc(coordCount_));
+                auto* steps = static_cast<uint8_t*>(platform::alloc(static_cast<size_t>(coordCount_) * 3));
                 if (grown && steps) {
                     if (keptIdx_) platform::free(keptIdx_);
-                    if (keptStep_) platform::free(keptStep_);
+                    if (keptPos_) platform::free(keptPos_);
                     keptIdx_ = grown;
-                    keptStep_ = steps;
+                    keptPos_ = steps;
                     keptIdxCap_ = coordCount_;
                     keptIdxAllocFailed_ = false;
                     publishHeapBytes();   // the index cache grew: refresh the memory readout
@@ -318,11 +324,13 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
                 if (x % p->s != 0 || y % p->s != 0 || z % p->s != 0) return;
                 PreviewDriver* self = p->self;
                 if (self->keptIdx_ && self->keptCount_ < self->keptIdxCap_) {
-                    // The coarsest lattice this light sits on: every axis divisible by the step.
-                    uint8_t step = 1;
-                    for (uint8_t t = 2; t <= 64; t++)
-                        if (x % t == 0 && y % t == 0 && z % t == 0) step = t;
-                    if (self->keptStep_) self->keptStep_[self->keptCount_] = step;
+                    // Bytes, like the coord table: a stride is at most 64.
+                    if (self->keptPos_) {
+                        const size_t at = static_cast<size_t>(self->keptCount_) * 3;
+                        self->keptPos_[at + 0] = static_cast<uint8_t>(x);
+                        self->keptPos_[at + 1] = static_cast<uint8_t>(y);
+                        self->keptPos_[at + 2] = static_cast<uint8_t>(z);
+                    }
                     self->keptIdx_[self->keptCount_++] = idx;
                 }
                 p->emit(x, y, z);
@@ -465,7 +473,7 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
         } else if (keptIdx_ && keptCount_ == coordCount_) {
             // The cached index map, filtered by the frame's stride: a light survives when that stride divides the coarsest lattice it sits on.
             for (nrOfLightsType k = 0; k < keptCount_; k++)
-                if (!keptStep_ || s == previewStride_ || keptStep_[k] % s == 0) col.emit(keptIdx_[k]);
+                if (!keptPos_ || s == previewStride_ || keptAtStride(k, s)) col.emit(keptIdx_[k]);
         } else {
             // The alloc-miss fallback: the full lattice walk, at the same stride as the table's.
             struct Skip { ColCtx* col; nrOfLightsType s; } sk{&col, s};
@@ -485,7 +493,7 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
     void freePreviewBuffers() {
         if (broadcaster_) broadcaster_->cancelBufferedSend();
         if (keptIdx_) { platform::free(keptIdx_); keptIdx_ = nullptr; keptIdxCap_ = 0; keptCount_ = 0; }
-        if (keptStep_) { platform::free(keptStep_); keptStep_ = nullptr; }
+        if (keptPos_) { platform::free(keptPos_); keptPos_ = nullptr; }
         if (staging_) { platform::free(staging_); staging_ = nullptr; stagingCap_ = 0; }
         publishHeapBytes();
     }
@@ -534,14 +542,15 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
     bool keptIdxAllocFailed_ = false;     // index cache couldn't allocate → gather walks per frame
     nrOfLightsType* keptIdx_ = nullptr;   // sparse layouts: kept lights' buffer indices, coord-table order
     nrOfLightsType keptIdxCap_ = 0, keptCount_ = 0;
-    // The coarsest stride each kept light survives, so a coarser frame filters the cache rather than re-walking placeLights (~8 ms at 12K lights, per frame).
-    uint8_t* keptStep_ = nullptr;
+    // Each kept light's (x,y,z): a coarser frame applies the device's own modulo test rather than re-walking placeLights (~8 ms at 12K lights, per frame).
+    uint8_t* keptPos_ = nullptr;
 
 protected:
     /// This driver's heap: the base scratch plus the kept-index cache, for the memory readout.
     size_t driverHeapBytes() const override {
         return DriverBase::driverHeapBytes()
              + static_cast<size_t>(keptIdxCap_) * sizeof(nrOfLightsType)
+             + static_cast<size_t>(keptIdxCap_) * 3   // keptPos_, three bytes per kept light
              + stagingCap_;
     }
 
diff --git a/src/ui/app.js b/src/ui/app.js
index d49f828a..1c7cf322 100644
--- a/src/ui/app.js
+++ b/src/ui/app.js
@@ -2369,7 +2369,7 @@ function formatStats(mod) {
     return head + statusChip;
 }
 function formatStatsTitle(mod) {
-    // The tooltip carries the OTHER unit, so both readings are available without clicking: the
+    // The tooltip carries the OTHER unit, so both readings are available without the gesture: the
     // card shows one, hovering says what it is in the other. A per-module fps is named as what it
     // is, a rate this module alone could sustain, because read as a frame rate it is a claim about
     // the whole pipeline made from one part of it.
@@ -2379,9 +2379,9 @@ function formatStatsTitle(mod) {
         const other = timingMode === "fps"
             ? (us < 1000 ? us + " µs" : (us / 1000).toFixed(2) + " ms") + " per tick"
             : fps.toLocaleString() + " fps if this module ran alone";
-        return other + "  ·  click to toggle";
+        return other + (timingMode === "fps" ? "  ·  release for time" : "  ·  hold for fps");
     }
-    return "Click to toggle fps/ms";
+    return "Hold to view fps";
 }
 
 function createActionButtons(mod) {
diff --git a/src/ui/preview3d.js b/src/ui/preview3d.js
index d897d315..9fa52a0c 100644
--- a/src/ui/preview3d.js
+++ b/src/ui/preview3d.js
@@ -692,12 +692,16 @@ function upsampleMap(stride) {
         const i = kept[k];
         slot.set(previewPos_[i * 3] + "," + previewPos_[i * 3 + 1] + "," + previewPos_[i * 3 + 2], k);
     }
+    // A block with no kept light is normal on a sparse layout: the lattice point its leader would
+    // occupy simply holds no light. Such a point draws DARK rather than borrowing colour 0, which
+    // is a real light elsewhere in the frame and would paint the wrong pixel.
+    const UNMAPPED = 0xffffffff;
     for (let i = 0; i < n; i++) {
         const bxq = previewPos_[i * 3] - (previewPos_[i * 3] % stride);
         const byq = previewPos_[i * 3 + 1] - (previewPos_[i * 3 + 1] % stride);
         const bzq = previewPos_[i * 3 + 2] - (previewPos_[i * 3 + 2] % stride);
         const k = slot.get(bxq + "," + byq + "," + bzq);
-        map[i] = k === undefined ? 0 : k;
+        map[i] = k === undefined ? UNMAPPED : k;
     }
     upsampleCache_ = new Map([[key, map]]);   // one epoch, one stride in flight: keep the last
     return map;
@@ -747,7 +751,11 @@ function renderPreviewFrame(view, buf) {
     if (up) {
         const full = new Uint8Array(previewCoordCount_ * 3);
         for (let i = 0; i < previewCoordCount_; i++) {
-            const src = up[i] * 3;
+            const k = up[i];
+            // Out of range means the frame and the map disagree (a mid-rebuild frame): draw dark
+            // rather than read past the body, which would repeat a stale colour.
+            if (k >= count) continue;
+            const src = k * 3;
             full[i * 3] = rgb[src]; full[i * 3 + 1] = rgb[src + 1]; full[i * 3 + 2] = rgb[src + 2];
         }
         previewRgb_ = full;
diff --git a/test/js/installer-s31-webflash.test.mjs b/test/js/installer-s31-webflash.test.mjs
index 3d60136f..ca349197 100644
--- a/test/js/installer-s31-webflash.test.mjs
+++ b/test/js/installer-s31-webflash.test.mjs
@@ -1,18 +1,13 @@
-// S31 web-flash guard contract — the installer special-cases chips esptool-js
-// (the browser flasher) can't handle, so a connect-flash failure shows "flash via
-// the CLI" guidance instead of a bare timeout. The guard keys on a chip-family
-// string (install.js `WEB_FLASH_UNSUPPORTED_CHIPS`); that string must match the
-// `chip` a board in deviceModels.json actually reports, or the guidance never fires.
+// Web-flash guard contract — the installer special-cases chips esptool-js (the browser flasher)
+// cannot handle, so a connect-flash failure shows "flash via the CLI" guidance instead of a bare
+// timeout. The guard keys on a chip-family string (install.js `WEB_FLASH_UNSUPPORTED_CHIPS`); that
+// string must match the `chip` a board in deviceModels.json actually reports, or it never fires.
 //
-// Background: the ESP32-S31's ROM magic collides with the classic ESP32's, and
-// esptool-js has only a magic table (no S31 secondary detection), so a browser
-// flash can't work safely — CLI flashing (esptool.py) is the path. See
-// install.js's WEB_FLASH_UNSUPPORTED_CHIPS comment + docs/backlog/backlog-core.md.
-//
-// This test pins two things so the guard can't silently break:
-//   1. install.js declares ESP32-S31 in WEB_FLASH_UNSUPPORTED_CHIPS, and
-//   2. a deviceModels.json board reports chip "ESP32-S31" (the exact guard string),
-// so the connect-flash message logic actually triggers for the S31 board.
+// The set is EMPTY today: esptool-js 0.7.0 added the ESP32-S31 target and chip-id detection
+// (GET_SECURITY_INFO), so every chip projectMM ships is browser-flashable. The guard stays because
+// a new chip lands here before esptool-js knows it, and these tests pin the MECHANISM rather than
+// any one entry: whatever the set holds must be a chip a catalog board reports, and the guidance
+// must stay scoped to the connect-flash stage.
 //
 // Run: `node --test test/js`.
 
@@ -32,16 +27,30 @@ const boards = JSON.parse(
 // The chip families install.js flags as not-browser-flashable. Parsed from the
 // source so the test reads the real declaration (no hard-coded copy that could drift).
 function unsupportedChips() {
-    const m = installJs.match(/WEB_FLASH_UNSUPPORTED_CHIPS\s*=\s*new Set\(\[([^\]]*)\]\)/);
+    // `new Set()` (empty) and `new Set([...])` are both valid declarations: the set is expected to
+    // be empty whenever esptool-js covers every chip we ship.
+    const m = installJs.match(/WEB_FLASH_UNSUPPORTED_CHIPS\s*=\s*new Set\((?:\[([^\]]*)\])?\)/);
     assert.ok(m, "install.js must declare WEB_FLASH_UNSUPPORTED_CHIPS as a Set literal");
-    return m[1].split(",").map((s) => s.trim().replace(/^["']|["']$/g, "")).filter(Boolean);
+    return (m[1] || "").split(",").map((s) => s.trim().replace(/^["']|["']$/g, "")).filter(Boolean);
 }
 
-test("install.js flags ESP32-S31 as not browser-flashable", () => {
+test("the S31 is browser-flashable: esptool-js 0.7.0 knows the chip", () => {
     assert.ok(
-        unsupportedChips().includes("ESP32-S31"),
-        "WEB_FLASH_UNSUPPORTED_CHIPS must include ESP32-S31 (esptool-js has no safe S31 support)"
+        !unsupportedChips().includes("ESP32-S31"),
+        "esptool-js 0.7.0 ships the ESP32-S31 target and chip-id detection, so the S31 no longer " +
+        "needs the CLI-only guard"
     );
+    // The pin is what makes that true: an older esptool-js has no S31 target, and its magic table
+    // would mis-identify the S31 as a classic ESP32 and flash the wrong stub.
+    const orchestrator = readFileSync(join(ROOT, "mooninstaller", "install-orchestrator.js"), "utf8");
+    const pin = orchestrator.match(/ESPTOOL_JS_VERSION\s*=\s*"(\d+)\.(\d+)\.(\d+)"/);
+    assert.ok(pin, "install-orchestrator.js must pin ESPTOOL_JS_VERSION");
+    const [maj, min] = [Number(pin[1]), Number(pin[2])];
+    assert.ok(maj > 0 || min >= 7,
+        `esptool-js must be >= 0.7.0 for S31 support, pinned ${pin[0]}`);
+    // The import must load the SAME version the footer credit shows.
+    assert.ok(orchestrator.includes(`esptool-js@${pin[1]}.${pin[2]}.${pin[3]}/bundle.js`),
+        "the unpkg import must match ESPTOOL_JS_VERSION");
 });
 
 test("the guard only fires on the connect-flash stage", () => {
diff --git a/test/unit/light/unit_PreviewDriver.cpp b/test/unit/light/unit_PreviewDriver.cpp
index 927e79f3..17674341 100644
--- a/test/unit/light/unit_PreviewDriver.cpp
+++ b/test/unit/light/unit_PreviewDriver.cpp
@@ -13,15 +13,15 @@
 #include <vector>
 #include <cstring>
 
-// PreviewDriver streams a true-shape point list: a one-time 0x03 coordinate table (positions of the real lights) + per-frame 0x02 RGB indexed by light. These tests pin: the table carries exactly lightCount positions (sphere → its shell count, NOT the bounding box), the per-frame RGB count matches, and a large layout is index-downsampled (stride > 1) to fit the send-buffer cap.
+// PreviewDriver streams a 0x03 table of the real lights' positions, then per-frame 0x02 RGB indexed by light. Pinned here: the table carries exactly lightCount positions (a sphere's shell, not its box), the frame agrees, and a large layout downsamples.
 
 namespace {
 
-// Captures the two preview message types so tests can inspect them. Every message arrives through the ONE resumable send (sendBufferedFrame) as header ++ body, classified by the type byte: 0x03 tables (11-byte header, epoch at [10]) into lastCoord, 0x02 frames (9-byte header, epoch at [7], drops at [8]) into lastFrame. dropCoord/acceptNext make a send report "slot busy" (false) to drive the request-retry and drop-counting paths.
+// Captures both message types. Each rides the one resumable send as header ++ body, by type byte: 0x03 tables (11-byte header, epoch [10]) to lastCoord, 0x02 frames (9-byte, epoch [7], drops [8]) to lastFrame. dropCoord/acceptNext report "slot busy" to drive the retry and drop paths.
 // -Wnon-virtual-dtor: BinaryBroadcaster's own destructor is protected and non-virtual on
 // purpose ("not owned through this interface"), so no code can delete through a base pointer.
-// This double is a stack local in every test, never owned polymorphically, and it cannot copy the base's protected-destructor trick, because that would forbid the stack construction the tests rely on. Scoped to this one type.
-// #ifndef _MSC_VER: `#pragma GCC` is an unknown pragma to MSVC (C4068), and the Windows build runs /WX, so an unguarded one fails it. MSVC has no -Wnon-virtual-dtor equivalent to silence, so excluding it there is complete, not a workaround. Clang understands `#pragma GCC`.
+// This double is a stack local in every test, never owned polymorphically, and cannot copy the base's protected-destructor trick: that would forbid the stack construction the tests rely on.
+// #ifndef _MSC_VER: `#pragma GCC` is unknown to MSVC (C4068) and /WX fails an unguarded one. MSVC has no -Wnon-virtual-dtor to silence, so excluding it there is complete. Clang understands it.
 #ifndef _MSC_VER
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wnon-virtual-dtor"
@@ -553,3 +553,27 @@ TEST_CASE("PreviewDriver strides a mapped layout's colors without re-walking the
     CHECK(rig.cap.frameCount() < rig.cap.coordCount());
     CHECK(rig.cap.frameBodyLights() == rig.cap.frameCount());
 }
+
+// Caching one "coarsest divisor" per light answers the stride test only while it IS the gcd: a loop capped at 64 truncated (65,65,65) to 13, so stride 5 dropped it.
+TEST_CASE("PreviewDriver keeps a sparse light whose axes share a factor above the cache's cap") {
+    mm::SphereLayout sph;
+    sph.radius = 64;                                  // its shell reaches coordinates gcd > 64
+    PreviewRig rig(&sph);
+    rig.cap.ask(5);
+    mm::platform::setTestNowMs(2000); rig.preview->tick();
+    mm::platform::setTestNowMs(0);
+    rig.produce();
+    CHECK(rig.cap.frameStride() == 5);
+    CHECK(rig.cap.frameBodyLights() == rig.cap.frameCount());
+
+    // The truth, counted independently: every shell light whose axes are all divisible by 5.
+    mm::SphereLayout ref;
+    ref.radius = 64;
+    struct Ctx { int keep; } cx{0};
+    ref.placeLights(mm::CoordSink{[](void* c, mm::nrOfLightsType, mm::lengthType x,
+                                     mm::lengthType y, mm::lengthType z) {
+        if (x % 5 == 0 && y % 5 == 0 && z % 5 == 0) static_cast<Ctx*>(c)->keep++;
+    }, nullptr, &cx});
+    REQUIRE(cx.keep > 0);
+    CHECK(rig.cap.frameCount() == cx.keep);
+}

From d508b4bada5ac58e651dcae19cca7218552bc81a Mon Sep 17 00:00:00 2001
From: ewowi <ewowi@icloud.com>
Date: Mon, 21 Sep 2026 23:46:20 +0200
Subject: [PATCH 5/5] Filter the preview on the bytes the browser holds
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The preview's kept-light cache stored raw coordinates while the table ships scaled ones, so past a 255 extent the device and the browser filtered different lattices and every colour landed on the wrong light. Beams were validated against the colour stride rather than the table's, which dropped them all on a slow link.

KPI: 16384lights | Desktop:1924KB | tick:125/96/354/2/6/124/258/21/1242/3/255/74/227/16/25/5/1528/124/21/6/146/47/4us(FPS:8000/10416/2824/500000/166666/8064/3875/47619/805/333333/3921/13513/4405/62500/40000/200000/654/8064/47619/166666/6849/21276/250000) | src:270(68904) | test:206(45223) | lizard:267w

**Light domain**
- `keptPos_` stores `scaleAxis(x)`, exactly what the coordinate table emits. A layout wider than 255 scales to v*255/extent on the wire, and the browser runs its modulo over those bytes: a 600-wide layout kept 300 lights device-side and 303 browser-side, so every colour past the first named the wrong light.
- `frameStride()` coarsens only where the kept-light cache can answer it. On the alloc-miss path the count and the gather filtered differently, so the header advertised more lights than the body carried and the browser rejected every frame: a blank preview on the degraded path, with no drop report to explain it.
- `ParallelLedDriver` carries an `@card`, so its screenshot reaches its generated page like every other driver's.
- The `0x52` docstring no longer advertises a stride byte the handler ignores.

**UI**
- An aim frame is validated against the TABLE's stride, which is what it is gathered at. Comparing the colour frame's, which the link moves independently, discarded every beam once the link coarsened.
- `activateTable`'s fallback picks the finest cached table rather than whichever arrived first: a coarser base cannot express a finer frame.

**Scripts/MoonDeck**
- `check_docgen` reads indentation from the raw line before stripping it. The two indent entries in the skip list could never match a stripped line, so an indented block could raise a blocking hard-wrap finding.

**Docs/CI**
- MIGRATING records `expertMode` becoming `mode`: a persisted control renamed and retyped, costing a re-set of one UI preference.

**Reviews**
- 👾 The kept-light cache filtered a different coordinate space than the browser → stores what the table sends, reproduced numerically at 300 vs 303 lights.
- 👾 The alloc-miss fallback advertised more lights than it sent → coarsens only when the cache can answer.
- 👾 A comment still described the divisor cache that a previous commit removed → rewritten.
- 👾 `ParallelLedDriver.png` reached no generated page → `@card` added.
- 👾 The table-cache fallback took the oldest table → takes the finest.
- 👾 `expertMode` → `mode` had no MIGRATING entry → added.
- 🐇 Indentation checked after stripping → checked on the raw line, control-checked both ways.
- 🐇 A request to widen `keptPos_` to `lengthType` → skipped: the wire format is one byte per axis, so widening would store what the browser never receives and reintroduce the divergence above.
- 🐇 A repo-health delta → generated output, rewritten by collect_kpi rather than edited.

Not covered by a test: the alloc-miss guard. Forcing an allocation failure needs a platform hook that does not exist, so the one-line guard is verified by reading rather than by a case.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 docs/reference/MIGRATING.md                   |  8 +++++
 docs/reference/metrics/repo-health.json       | 32 ++++++++---------
 docs/reference/metrics/repo-health.md         | 36 +++++++++----------
 moondeck/check/check_docgen.py                | 10 ++++--
 src/light/drivers/ParallelLedDriver.h         |  1 +
 src/light/drivers/PreviewDriver.h             | 17 +++++----
 src/ui/preview3d.js                           | 10 ++++--
 .../scenario_MoonModule_control_change.json   | 12 +++----
 .../light/scenario_Audio_mutation.json        | 16 ++++-----
 test/scenarios/light/scenario_Aurora_fps.json | 30 ++++++++--------
 .../light/scenario_Driver_mutation.json       | 10 +++---
 .../light/scenario_Effects_composition.json   |  2 +-
 .../light/scenario_Fields_polar_lut.json      | 30 ++++++++--------
 .../light/scenario_Fluid_solver.json          | 30 ++++++++--------
 .../light/scenario_GridLayout_resize.json     |  6 ++--
 .../light/scenario_Layer_base_pipeline.json   |  2 +-
 .../light/scenario_Layer_memory_1to1.json     |  2 +-
 .../light/scenario_Layouts_mutation.json      |  8 ++---
 .../scenario_MoonLiveEffect_livescript.json   | 24 ++++++-------
 .../light/scenario_MoonLive_pipeline.json     | 10 +++---
 .../scenario_MultiplyModifier_memory_lut.json |  2 +-
 .../scenario_MultiplyModifier_pipeline.json   |  2 +-
 .../light/scenario_Trails_ladder.json         | 16 ++++-----
 .../light/scenario_modifier_chain.json        | 10 +++---
 .../light/scenario_modifier_swap.json         |  6 ++--
 test/scenarios/light/scenario_perf_full.json  | 26 +++++++-------
 test/scenarios/light/scenario_perf_light.json |  4 +--
 .../light/scenario_peripheral_grid_sweep.json | 34 +++++++++---------
 .../light/scenario_peripheral_switch.json     |  4 +--
 29 files changed, 211 insertions(+), 189 deletions(-)

diff --git a/docs/reference/MIGRATING.md b/docs/reference/MIGRATING.md
index 86fc33be..c29b5c45 100644
--- a/docs/reference/MIGRATING.md
+++ b/docs/reference/MIGRATING.md
@@ -24,6 +24,14 @@ projectMM ships **no migration code**: the persistence layer is robust by defaul
 
 ## Unreleased (`next-iteration`)
 
+### System: `expertMode` became `mode`, with three levels
+
+**Action: re-set a control, and only if you had expert mode on.**
+
+The switch that revealed advanced controls is now a three-way select: `user`, `expert` (🎚️) and `developer` (🔧), each level showing what the one below it shows. A saved `expertMode` no longer matches a control and is dropped, so a device comes up in `user` mode whatever it held before. Pick the level you want again on the System card.
+
+The old flag could only say "show more" or "show less", which left diagnostics that mean nothing without the source sitting beside the controls a light show is built from.
+
 ### Documentation: every page moved to its Diátaxis folder
 
 **Action: nothing on a device. Update a bookmark to a documentation page.**
diff --git a/docs/reference/metrics/repo-health.json b/docs/reference/metrics/repo-health.json
index d27ed1b0..bbd75663 100644
--- a/docs/reference/metrics/repo-health.json
+++ b/docs/reference/metrics/repo-health.json
@@ -1,13 +1,13 @@
 {
-  "commit": "8f50ae79",
+  "commit": "eaa904dc",
   "flash": {
     "esp32s3-n16r8": 2134032,
     "desktop": 1970712,
-    "esp32": 2084912,
-    "esp32p4rev1-eth": 2034592,
+    "esp32": 2085360,
+    "esp32p4rev1-eth": 2035152,
     "esp32p4rev1-eth-wifi": 2284640,
     "esp32s3-n8r8": 2087168,
-    "esp32s31": 2427632,
+    "esp32s31": 2428080,
     "esp32-16mb": 2060368,
     "esp32-eth": 1681040,
     "esp32-wrover": 1843760,
@@ -31,8 +31,8 @@
   },
   "perf": {
     "desktop": {
-      "tick_us": 121,
-      "fps": 8264,
+      "tick_us": 125,
+      "fps": 8000,
       "scenario_p50": {
         "Layer_base_pipeline": {
           "p50": 70,
@@ -197,8 +197,8 @@
       },
       "Aurora_fps": {
         "desktop-macos": {
-          "p50": 1521,
-          "p95": 1947,
+          "p50": 1522,
+          "p95": 1951,
           "n": 32,
           "last": "2026-09-21"
         }
@@ -512,7 +512,7 @@
       "modifier_chain": {
         "desktop-macos": {
           "p50": 43,
-          "p95": 46,
+          "p95": 47,
           "n": 32,
           "last": "2026-09-21"
         },
@@ -699,11 +699,11 @@
   },
   "loc": {
     "core": 22010,
-    "light": 29961,
+    "light": 29965,
     "platform": 16346,
-    "ui": 11318,
+    "ui": 11324,
     "test": 56401,
-    "moondeck": 27237
+    "moondeck": 27241
   },
   "comments": {
     "core": {
@@ -711,7 +711,7 @@
       "ratio": 0.28
     },
     "light": {
-      "lines": 7270,
+      "lines": 7272,
       "ratio": 0.272
     },
     "platform": {
@@ -719,7 +719,7 @@
       "ratio": 0.238
     },
     "ui": {
-      "lines": 3376,
+      "lines": 3381,
       "ratio": 0.315
     },
     "test": {
@@ -727,7 +727,7 @@
       "ratio": 0.131
     },
     "moondeck": {
-      "lines": 4572,
+      "lines": 4575,
       "ratio": 0.191
     }
   },
@@ -737,7 +737,7 @@
   },
   "docs": {
     "md_files": 138,
-    "md_lines": 28282,
+    "md_lines": 28290,
     "plans_files": 31,
     "backlog_lines": 3034,
     "lessons_lines": 526,
diff --git a/docs/reference/metrics/repo-health.md b/docs/reference/metrics/repo-health.md
index b9c21c4a..6b6219e2 100644
--- a/docs/reference/metrics/repo-health.md
+++ b/docs/reference/metrics/repo-health.md
@@ -1,6 +1,6 @@
 # Repo health
 
-Measured at `8f50ae79`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
+Measured at `eaa904dc`. Generated by [`moondeck/check/repo_health.py`](../../../moondeck/check/repo_health.py) on every KPI-gate run. **Do not edit by hand.**
 
 Current state only; the trend is this file's git history (`git log -p docs/reference/metrics/repo-health.md`). Nothing here fails a build: the numbers make growth visible, the judgment stays human.
 
@@ -8,19 +8,19 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Flash | Capacity | Used | Built |
 |---|---:|---:|---:|:--:|
-| desktop | 1,925 KB (+48 B) ⚠ | - | - | yes |
-| esp32 | 2,036 KB (+240 B) ⚠ | 2,496 KB | 82% | yes |
+| desktop | 1,925 KB | - | - | yes |
+| esp32 | 2,036 KB (+448 B) ⚠ | 2,496 KB | 82% | yes |
 | esp32-16mb | 2,012 KB | 4,096 KB | 49% | **STALE 12d** |
 | esp32-eth | 1,642 KB | 2,496 KB | 66% | **STALE 10d** |
 | esp32-pico | 2,058 KB | 3,072 KB | 67% | **STALE 12d** |
 | esp32-wrover | 1,801 KB | - | - | carried (age?) |
-| esp32p4rev1-eth | 1,987 KB (+304 B) ⚠ | 4,096 KB | 49% | yes |
+| esp32p4rev1-eth | 1,987 KB (+560 B) ⚠ | 4,096 KB | 49% | yes |
 | esp32p4rev1-eth-wifi | 2,231 KB | 4,096 KB | 54% | **STALE 13d** |
 | esp32p4rev3-eth | 1,605 KB | - | - | carried (age?) |
-| esp32s3-n16r8 | 2,084 KB | 4,096 KB | 51% | yes |
+| esp32s3-n16r8 | 2,084 KB | 4,096 KB | 51% | carried 0d |
 | esp32s3-n8r8 | 2,038 KB | 3,072 KB | 66% | **STALE 13d** |
 | esp32s3-zero | 1,977 KB | 2,496 KB | 79% | **STALE 13d** |
-| esp32s31 | 2,371 KB (+192 B) ⚠ | 4,096 KB | 58% | yes |
+| esp32s31 | 2,371 KB (+448 B) ⚠ | 4,096 KB | 58% | yes |
 | qemu | 1,351 KB | - | - | carried (age?) |
 
 `Built: yes` was measured this run. `carried (age?)` was not rebuilt either and predates this record, so its age is unknown: it dates itself on the next build. `carried Nd` was NOT rebuilt and its number is N days old, so an absent delta says nothing about the change. **STALE** marks a carry older than 7 days: the number has gone unchecked long enough that growth will surface later as one jump, blamed on whichever commit happens to rebuild that target. `Used` is against the app slot in the firmware's own partition table.
@@ -29,7 +29,7 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 
 | Target | Tick | FPS |
 |---|---:|---:|
-| desktop | 121 µs (−3 µs) ✓ | 8,264 (+200) ✓ |
+| desktop | 125 µs (+4 µs) ⚠ | 8,000 (−264) ⚠ |
 | esp32 | 8,354 µs | 119 |
 
 ### Scenario tick by target (p50 of each sample window)
@@ -37,7 +37,7 @@ Current state only; the trend is this file's git history (`git log -p docs/refer
 | Scenario | desktop-macos | desktop-windows | esp32 | esp32s3-n16r8 | esp32p4rev1-eth | esp32s31 | esp32-eth | esp32-eth-wifi | unknown |
 |---|---|---|---|---|---|---|---|---|---|
 | Audio_mutation | 22 | 40 ? | 13,152 | 47 ? | - | - | - | - | - |
-| Aurora_fps | 1,521 | - | - | - | - | - | - | - | - |
+| Aurora_fps | 1,522 (+1) ⚠ | - | - | - | - | - | - | - | - |
 | Driver_mutation | 20 | 42 ? | 12,812 | 39 ? | - | - | - | - | - |
 | Effects_composition | 144 | 549 ? | - | - | - | - | - | - | - |
 | Fields_polar_lut | 1,263 | - | - | - | - | - | - | - | - |
@@ -82,24 +82,24 @@ These build a bare pipeline with no optional modules, so a change here is a chan
 | Area | Lines | Comments | Comment share |
 |---|---:|---:|---:|
 | core | 22,010 | 5,579 | 28.0 % |
-| light | 29,961 (+9) ⚠ | 7,270 | 27.2 % |
+| light | 29,965 (+4) ⚠ | 7,272 | 27.2 % |
 | platform | 16,346 | 3,481 | 23.8 % |
-| ui | 11,318 (+8) ⚠ | 3,376 | 31.5 % |
-| test | 56,401 (+24) ⚠ | 6,362 | 13.1 % |
-| moondeck | 27,237 (+2) ⚠ | 4,572 | 19.1 % |
+| ui | 11,324 (+6) ⚠ | 3,381 | 31.5 % |
+| test | 56,401 | 6,362 | 13.1 % |
+| moondeck | 27,241 (+4) ⚠ | 4,575 | 19.1 % |
 
 ## Tests
 
 | Kind | Count |
 |---|---:|
-| unit cases | 2,071 (+1) ✓ |
+| unit cases | 2,071 |
 | scenarios | 27 |
 
 ## Complexity
 
 | Metric | Value |
 |---|---:|
-| functions | 3,631 (+1) ✓ |
+| functions | 3,631 |
 | over threshold | 267 |
 | worst CCN | 128 |
 
@@ -108,9 +108,9 @@ These build a bare pipeline with no optional modules, so a change here is a chan
 | Metric | Value |
 |---|---:|
 | markdown files | 138 |
-| markdown lines | 28,282 (+17) ⚠ |
+| markdown lines | 28,290 (+8) ⚠ |
 | plan files | 31 |
-| backlog lines | 3,034 (+9) ⚠ |
-| lessons lines | 526 (+8) ⚠ |
-| CLAUDE.md lines | 281 (+2) ⚠ |
+| backlog lines | 3,034 |
+| lessons lines | 526 |
+| CLAUDE.md lines | 281 |
 
diff --git a/moondeck/check/check_docgen.py b/moondeck/check/check_docgen.py
index efca65b5..835efa42 100755
--- a/moondeck/check/check_docgen.py
+++ b/moondeck/check/check_docgen.py
@@ -1358,7 +1358,8 @@ def _duplicate_group_ids():
 # forbid it for the same reason the header rule does: the editor soft-wraps, so a one-word edit
 # is a one-word diff rather than a reflowed paragraph. Checked over AUTHORED pages only, since a
 # generated one carries its source's shape and is fixed at the generator.
-_MD_SKIP_PREFIX = ("#", "|", "-", "*", ">", "<", "```", "    ", "\t", "!", "=", ":")
+# Indentation is tested on the RAW line above, since these are matched after stripping.
+_MD_SKIP_PREFIX = ("#", "|", "-", "*", ">", "<", "```", "!", "=", ":")
 _MD_NUMBERED = re.compile(r"^\d+[.)]\s")
 
 
@@ -1387,10 +1388,13 @@ def _md_hard_wraps(rel: str, text: str):
             continue
         if fenced or not stripped:
             continue
-        if stripped.startswith(_MD_SKIP_PREFIX) or _MD_NUMBERED.match(stripped):
+        # INDENTATION is read before stripping: an indented block is a listing or a fenceless code
+        # sample, not prose, and the stripped line can never carry the leading whitespace that says so.
+        if line.startswith(("    ", "\t")) or stripped.startswith(_MD_SKIP_PREFIX) or _MD_NUMBERED.match(stripped):
             continue
         nxt = lines[i + 1].strip()
-        if not nxt or nxt.startswith(_MD_SKIP_PREFIX) or _MD_NUMBERED.match(nxt):
+        if (not nxt or lines[i + 1].startswith(("    ", "\t"))
+                or nxt.startswith(_MD_SKIP_PREFIX) or _MD_NUMBERED.match(nxt)):
             continue
         # A line ending mid-sentence is the wrap; one ending a sentence is a deliberate break.
         if stripped[-1] not in ".!?:":
diff --git a/src/light/drivers/ParallelLedDriver.h b/src/light/drivers/ParallelLedDriver.h
index 82990bcd..aaf6b430 100644
--- a/src/light/drivers/ParallelLedDriver.h
+++ b/src/light/drivers/ParallelLedDriver.h
@@ -15,6 +15,7 @@ namespace mm {
 ///
 /// The whole frame is encoded up front and shipped as one autonomous transfer. So there is no CPU deadline while it is on the wire. The encode is a fused correct and transpose, per row (ParallelSlots.h). Vocabulary: strand, lane, slot, row, under
 /// @xref{terminology|More info → Terminology}.
+/// @card ParallelLedDriver.png
 ///
 /// @moreinfo
 ///
diff --git a/src/light/drivers/PreviewDriver.h b/src/light/drivers/PreviewDriver.h
index 801d754e..a963ce88 100644
--- a/src/light/drivers/PreviewDriver.h
+++ b/src/light/drivers/PreviewDriver.h
@@ -28,7 +28,7 @@ namespace mm {
 ///      [0x02][count:u32][stride:u16][epoch:u8][drops:u8][(r,g,b) x count]
 /// 0x04 per-frame aim, only for a rig whose fixtures carry pan and tilt:
 ///      [0x04][count:u32][stride:u16][epoch:u8][reserved:u8][(pan,tilt):u8x2 x count]
-/// Client requests: [0x51][stride][fps] standing, [0x52][stride] one-shot table.
+/// Client requests: [0x51][stride][fps] standing, [0x52] one-shot table. The table's own stride is memory-derived, so the request carries none.
 /// ```
 /// --8<-- [end:wire-format]
 ///
@@ -81,6 +81,9 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
 
     /// The stride a color frame ships at: the link's pacing, never finer than the table's.
     nrOfLightsType frameStride() const {
+        // Coarsening needs the kept-light cache, or the count and the gather filter differently and the browser rejects every frame.
+        const bool canCoarsen = denseGrid() || (keptPos_ && keptCount_ == coordCount_);
+        if (!canCoarsen) return previewStride_;
         return downscale_ > previewStride_ ? downscale_ : previewStride_;
     }
 
@@ -324,12 +327,12 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
                 if (x % p->s != 0 || y % p->s != 0 || z % p->s != 0) return;
                 PreviewDriver* self = p->self;
                 if (self->keptIdx_ && self->keptCount_ < self->keptIdxCap_) {
-                    // Bytes, like the coord table: a stride is at most 64.
+                    // SCALED, as the coord table emits them: past a 255 extent it carries v*255/extent, and the browser runs its modulo over those bytes.
                     if (self->keptPos_) {
                         const size_t at = static_cast<size_t>(self->keptCount_) * 3;
-                        self->keptPos_[at + 0] = static_cast<uint8_t>(x);
-                        self->keptPos_[at + 1] = static_cast<uint8_t>(y);
-                        self->keptPos_[at + 2] = static_cast<uint8_t>(z);
+                        self->keptPos_[at + 0] = self->scaleAxis(x);
+                        self->keptPos_[at + 1] = self->scaleAxis(y);
+                        self->keptPos_[at + 2] = self->scaleAxis(z);
                     }
                     self->keptIdx_[self->keptCount_++] = idx;
                 }
@@ -471,9 +474,9 @@ class PreviewDriver : public DriverBase, public BinaryBroadcaster::ClientMessage
                         col.emit(static_cast<nrOfLightsType>(static_cast<size_t>(z) * H * W
                                                              + static_cast<size_t>(y) * W + x));
         } else if (keptIdx_ && keptCount_ == coordCount_) {
-            // The cached index map, filtered by the frame's stride: a light survives when that stride divides the coarsest lattice it sits on.
+            // The cached index map, filtered by the frame's stride on the same bytes the browser holds.
             for (nrOfLightsType k = 0; k < keptCount_; k++)
-                if (!keptPos_ || s == previewStride_ || keptAtStride(k, s)) col.emit(keptIdx_[k]);
+                if (!keptPos_ || s == previewStride_ || keptAtStride(k, s)) col.emit(keptIdx_[k]);   // the stride test, on the bytes the browser has
         } else {
             // The alloc-miss fallback: the full lattice walk, at the same stride as the table's.
             struct Skip { ColCtx* col; nrOfLightsType s; } sk{&col, s};
diff --git a/src/ui/preview3d.js b/src/ui/preview3d.js
index 9fa52a0c..a8cf60b0 100644
--- a/src/ui/preview3d.js
+++ b/src/ui/preview3d.js
@@ -650,9 +650,12 @@ function parsePreviewCoords(view, buf) {
 // asks the device for it: the pull model).
 function activateTable(epoch, stride) {
     // Any cached table for this epoch serves every stride: the colors are mapped onto it below.
+    // The FINEST one when several are cached, since a coarser base cannot express a finer frame:
+    // insertion order would have handed back whichever arrived first.
     const t = tableCache_.get(epoch + ":" + stride)
            || [...tableCache_.keys()].filter(k => k.startsWith(epoch + ":"))
-                .map(k => tableCache_.get(k))[0];
+                .map(k => tableCache_.get(k))
+                .sort((a, b) => (a.stride ?? 1) - (b.stride ?? 1))[0];
     if (!t) return false;
     previewCoords_ = t.coords;
     previewCoordCount_ = t.count;
@@ -1041,7 +1044,10 @@ function drawBeams(mvp) {
     if (!previewAim_ || !previewCoords_ || !beamProgram) return;
     // Stale aim: gathered against a table that is no longer active, so its indices name other
     // fixtures now. Skip until the next aim frame rather than draw beams from the wrong heads.
-    if (previewAimEpoch_ !== lastEpoch_ || previewAimStride_ !== previewStride_) return;
+    // Against the TABLE's stride, which is what the aim frame is gathered at: the color frame's
+    // own stride is the link's and moves independently, so comparing it dropped every beam on a
+    // slow link.
+    if (previewAimEpoch_ !== lastEpoch_ || previewAimStride_ !== previewTableStride_) return;
     const n = Math.min(previewAim_.length >> 1, previewCoords_.length / 3);
     if (n === 0) return;
 
diff --git a/test/scenarios/core/scenario_MoonModule_control_change.json b/test/scenarios/core/scenario_MoonModule_control_change.json
index c206031f..31217e14 100644
--- a/test/scenarios/core/scenario_MoonModule_control_change.json
+++ b/test/scenarios/core/scenario_MoonModule_control_change.json
@@ -122,7 +122,7 @@
             "min": 115,
             "max": 198,
             "n": 32,
-            "samples": [163, 194, 186, 133, 122, 196, 123, 194, 119, 121, 126, 126, 121, 123, 117, 194, 117, 123, 121, 124, 194, 124, 193, 191, 192, 198, 194, 142, 119, 115, 120, 119]
+            "samples": [194, 186, 133, 122, 196, 123, 194, 119, 121, 126, 126, 121, 123, 117, 194, 117, 123, 121, 124, 194, 124, 193, 191, 192, 198, 194, 142, 119, 115, 120, 119, 122]
           },
           "last_updated": "2026-09-21"
         },
@@ -304,7 +304,7 @@
             "min": 119,
             "max": 139,
             "n": 32,
-            "samples": [135, 131, 129, 131, 123, 132, 124, 130, 119, 122, 127, 139, 120, 122, 119, 130, 119, 128, 124, 121, 131, 139, 130, 131, 129, 132, 131, 127, 122, 120, 120, 119]
+            "samples": [131, 129, 131, 123, 132, 124, 130, 119, 122, 127, 139, 120, 122, 119, 130, 119, 128, 124, 121, 131, 139, 130, 131, 129, 132, 131, 127, 122, 120, 120, 119, 124]
           },
           "last_updated": "2026-09-21"
         },
@@ -486,7 +486,7 @@
             "min": 115,
             "max": 126,
             "n": 32,
-            "samples": [116, 118, 117, 120, 121, 120, 121, 115, 119, 125, 125, 118, 121, 121, 120, 117, 119, 123, 126, 120, 117, 123, 118, 116, 115, 117, 116, 116, 122, 121, 121, 120]
+            "samples": [118, 117, 120, 121, 120, 121, 115, 119, 125, 125, 118, 121, 121, 120, 117, 119, 123, 126, 120, 117, 123, 118, 116, 115, 117, 116, 116, 122, 121, 121, 120, 123]
           },
           "last_updated": "2026-09-21"
         },
@@ -671,12 +671,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 119,
+            "p50": 120,
             "p95": 125,
-            "min": 117,
+            "min": 118,
             "max": 128,
             "n": 32,
-            "samples": [117, 119, 119, 120, 121, 121, 119, 118, 119, 122, 125, 120, 120, 121, 120, 119, 119, 128, 125, 120, 119, 124, 118, 119, 118, 119, 120, 119, 119, 119, 120, 120]
+            "samples": [119, 119, 120, 121, 121, 119, 118, 119, 122, 125, 120, 120, 121, 120, 119, 119, 128, 125, 120, 119, 124, 118, 119, 118, 119, 120, 119, 119, 119, 120, 120, 123]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Audio_mutation.json b/test/scenarios/light/scenario_Audio_mutation.json
index 45b2044d..ad54d90e 100644
--- a/test/scenarios/light/scenario_Audio_mutation.json
+++ b/test/scenarios/light/scenario_Audio_mutation.json
@@ -109,7 +109,7 @@
             "min": 16,
             "max": 39,
             "n": 32,
-            "samples": [16, 16, 16, 20, 16, 20, 16, 19, 18, 21, 16, 16, 20, 16, 16, 17, 39, 17, 20, 20, 17, 16, 16, 16, 17, 16, 17, 17, 17, 20, 20, 20]
+            "samples": [16, 16, 20, 16, 20, 16, 19, 18, 21, 16, 16, 20, 16, 16, 17, 39, 17, 20, 20, 17, 16, 16, 16, 17, 16, 17, 17, 17, 20, 20, 20, 17]
           },
           "last_updated": "2026-09-21"
         },
@@ -206,7 +206,7 @@
             "min": 17,
             "max": 186,
             "n": 32,
-            "samples": [17, 25, 19, 19, 21, 18, 20, 21, 18, 20, 17, 43, 41, 27, 25, 23, 26, 36, 17, 20, 186, 21, 33, 21, 18, 26, 20, 18, 17, 39, 26, 20]
+            "samples": [25, 19, 19, 21, 18, 20, 21, 18, 20, 17, 43, 41, 27, 25, 23, 26, 36, 17, 20, 186, 21, 33, 21, 18, 26, 20, 18, 17, 39, 26, 20, 41]
           },
           "last_updated": "2026-09-21"
         },
@@ -315,12 +315,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 19,
+            "p50": 20,
             "p95": 48,
             "min": 17,
             "max": 109,
             "n": 32,
-            "samples": [18, 28, 18, 20, 17, 19, 19, 21, 19, 20, 17, 19, 34, 24, 19, 20, 30, 30, 18, 23, 109, 29, 32, 19, 17, 19, 19, 18, 17, 26, 48, 20]
+            "samples": [28, 18, 20, 17, 19, 19, 21, 19, 20, 17, 19, 34, 24, 19, 20, 30, 30, 18, 23, 109, 29, 32, 19, 17, 19, 19, 18, 17, 26, 48, 20, 36]
           },
           "last_updated": "2026-09-21"
         },
@@ -417,7 +417,7 @@
             "min": 18,
             "max": 61,
             "n": 32,
-            "samples": [19, 38, 22, 20, 18, 24, 24, 22, 23, 21, 22, 20, 61, 21, 21, 23, 34, 28, 22, 21, 21, 27, 28, 20, 19, 20, 21, 23, 18, 26, 29, 23]
+            "samples": [38, 22, 20, 18, 24, 24, 22, 23, 21, 22, 20, 61, 21, 21, 23, 34, 28, 22, 21, 21, 27, 28, 20, 19, 20, 21, 23, 18, 26, 29, 23, 25]
           },
           "last_updated": "2026-09-21"
         },
@@ -512,7 +512,7 @@
             "min": 18,
             "max": 50,
             "n": 32,
-            "samples": [18, 26, 21, 20, 18, 35, 20, 23, 21, 21, 22, 21, 22, 19, 19, 20, 21, 50, 22, 19, 19, 40, 21, 19, 21, 23, 21, 22, 21, 21, 19, 23]
+            "samples": [26, 21, 20, 18, 35, 20, 23, 21, 21, 22, 21, 22, 19, 19, 20, 21, 50, 22, 19, 19, 40, 21, 19, 21, 23, 21, 22, 21, 21, 19, 23, 19]
           },
           "last_updated": "2026-09-21"
         },
@@ -602,12 +602,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 20,
+            "p50": 19,
             "p95": 22,
             "min": 16,
             "max": 23,
             "n": 32,
-            "samples": [20, 18, 20, 20, 17, 21, 21, 19, 21, 20, 20, 20, 19, 17, 17, 18, 21, 23, 20, 22, 18, 19, 18, 17, 20, 19, 17, 20, 20, 19, 16, 20]
+            "samples": [18, 20, 20, 17, 21, 21, 19, 21, 20, 20, 20, 19, 17, 17, 18, 21, 23, 20, 22, 18, 19, 18, 17, 20, 19, 17, 20, 20, 19, 16, 20, 17]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Aurora_fps.json b/test/scenarios/light/scenario_Aurora_fps.json
index 9901b245..9c7eb04e 100644
--- a/test/scenarios/light/scenario_Aurora_fps.json
+++ b/test/scenarios/light/scenario_Aurora_fps.json
@@ -89,7 +89,7 @@
             "min": 819,
             "max": 1744,
             "n": 32,
-            "samples": [873, 821, 923, 837, 846, 890, 853, 875, 854, 851, 843, 919, 841, 1005, 1177, 852, 880, 907, 862, 839, 1744, 877, 929, 1001, 834, 899, 864, 835, 819, 836, 840, 862]
+            "samples": [821, 923, 837, 846, 890, 853, 875, 854, 851, 843, 919, 841, 1005, 1177, 852, 880, 907, 862, 839, 1744, 877, 929, 1001, 834, 899, 864, 835, 819, 836, 840, 862, 929]
           },
           "last_updated": "2026-09-21"
         }
@@ -114,12 +114,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 321,
+            "p50": 322,
             "p95": 478,
             "min": 308,
             "max": 936,
             "n": 32,
-            "samples": [312, 308, 322, 318, 317, 321, 321, 315, 326, 323, 320, 936, 318, 336, 320, 322, 331, 478, 332, 316, 444, 333, 346, 405, 315, 349, 349, 324, 309, 316, 316, 317]
+            "samples": [308, 322, 318, 317, 321, 321, 315, 326, 323, 320, 936, 318, 336, 320, 322, 331, 478, 332, 316, 444, 333, 346, 405, 315, 349, 349, 324, 309, 316, 316, 317, 324]
           },
           "last_updated": "2026-09-21"
         }
@@ -149,7 +149,7 @@
             "min": 180,
             "max": 391,
             "n": 32,
-            "samples": [181, 181, 189, 188, 189, 190, 191, 185, 189, 189, 195, 301, 186, 190, 201, 189, 196, 261, 194, 185, 373, 188, 197, 391, 184, 207, 186, 182, 180, 185, 185, 185]
+            "samples": [181, 189, 188, 189, 190, 191, 185, 189, 189, 195, 301, 186, 190, 201, 189, 196, 261, 194, 185, 373, 188, 197, 391, 184, 207, 186, 182, 180, 185, 185, 185, 189]
           },
           "last_updated": "2026-09-21"
         }
@@ -174,12 +174,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 551,
+            "p50": 552,
             "p95": 899,
             "min": 529,
             "max": 1141,
             "n": 32,
-            "samples": [532, 535, 559, 551, 546, 549, 549, 541, 607, 567, 719, 899, 547, 558, 552, 674, 584, 1141, 568, 543, 662, 559, 600, 642, 541, 579, 542, 540, 529, 546, 544, 546]
+            "samples": [535, 559, 551, 546, 549, 549, 541, 607, 567, 719, 899, 547, 558, 552, 674, 584, 1141, 568, 543, 662, 559, 600, 642, 541, 579, 542, 540, 529, 546, 544, 546, 556]
           },
           "last_updated": "2026-09-21"
         }
@@ -204,12 +204,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 940,
+            "p50": 943,
             "p95": 1265,
             "min": 910,
             "max": 1683,
             "n": 32,
-            "samples": [918, 915, 960, 1015, 940, 945, 949, 938, 934, 977, 1001, 1253, 939, 943, 940, 938, 978, 1683, 1000, 939, 1070, 981, 978, 1265, 931, 998, 932, 917, 910, 938, 939, 936]
+            "samples": [915, 960, 1015, 940, 945, 949, 938, 934, 977, 1001, 1253, 939, 943, 940, 938, 978, 1683, 1000, 939, 1070, 981, 978, 1265, 931, 998, 932, 917, 910, 938, 939, 936, 960]
           },
           "last_updated": "2026-09-21"
         }
@@ -234,12 +234,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 1458,
+            "p50": 1460,
             "p95": 1665,
             "min": 1407,
             "max": 1678,
             "n": 32,
-            "samples": [1417, 1410, 1461, 1454, 1453, 1462, 1460, 1448, 1443, 1470, 1502, 1486, 1465, 1458, 1460, 1443, 1498, 1555, 1527, 1446, 1503, 1517, 1665, 1678, 1440, 1475, 1440, 1414, 1407, 1445, 1450, 1447]
+            "samples": [1410, 1461, 1454, 1453, 1462, 1460, 1448, 1443, 1470, 1502, 1486, 1465, 1458, 1460, 1443, 1498, 1555, 1527, 1446, 1503, 1517, 1665, 1678, 1440, 1475, 1440, 1414, 1407, 1445, 1450, 1447, 1479]
           },
           "last_updated": "2026-09-21"
         }
@@ -264,12 +264,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 1521,
-            "p95": 1947,
+            "p50": 1522,
+            "p95": 1951,
             "min": 1453,
-            "max": 1951,
+            "max": 2061,
             "n": 32,
-            "samples": [1496, 1498, 1899, 1508, 1648, 1506, 1505, 1505, 1493, 1624, 1526, 1521, 1535, 1500, 1503, 1624, 1564, 1590, 1951, 1590, 1517, 1549, 1679, 1544, 1489, 1947, 1550, 1455, 1453, 1522, 1498, 1497]
+            "samples": [1498, 1899, 1508, 1648, 1506, 1505, 1505, 1493, 1624, 1526, 1521, 1535, 1500, 1503, 1624, 1564, 1590, 1951, 1590, 1517, 1549, 1679, 1544, 1489, 1947, 1550, 1455, 1453, 1522, 1498, 1497, 2061]
           },
           "last_updated": "2026-09-21"
         }
@@ -299,7 +299,7 @@
             "min": 1071,
             "max": 1828,
             "n": 32,
-            "samples": [1078, 1088, 1131, 1119, 1117, 1109, 1117, 1101, 1099, 1828, 1258, 1115, 1121, 1148, 1103, 1223, 1150, 1163, 1142, 1120, 1116, 1150, 1137, 1134, 1108, 1108, 1177, 1077, 1071, 1106, 1103, 1102]
+            "samples": [1088, 1131, 1119, 1117, 1109, 1117, 1101, 1099, 1828, 1258, 1115, 1121, 1148, 1103, 1223, 1150, 1163, 1142, 1120, 1116, 1150, 1137, 1134, 1108, 1108, 1177, 1077, 1071, 1106, 1103, 1102, 1145]
           },
           "last_updated": "2026-09-21"
         }
diff --git a/test/scenarios/light/scenario_Driver_mutation.json b/test/scenarios/light/scenario_Driver_mutation.json
index c0690939..70e33235 100644
--- a/test/scenarios/light/scenario_Driver_mutation.json
+++ b/test/scenarios/light/scenario_Driver_mutation.json
@@ -81,7 +81,7 @@
             "min": 16,
             "max": 46,
             "n": 32,
-            "samples": [21, 20, 16, 20, 20, 46, 23, 20, 19, 16, 20, 20, 17, 17, 17, 17, 19, 19, 19, 19, 23, 21, 30, 17, 20, 17, 17, 20, 16, 20, 20, 20]
+            "samples": [20, 16, 20, 20, 46, 23, 20, 19, 16, 20, 20, 17, 17, 17, 17, 19, 19, 19, 19, 23, 21, 30, 17, 20, 17, 17, 20, 16, 20, 20, 20, 21]
           },
           "last_updated": "2026-09-21"
         },
@@ -178,7 +178,7 @@
             "min": 16,
             "max": 38,
             "n": 32,
-            "samples": [20, 20, 16, 20, 19, 38, 19, 20, 21, 16, 20, 21, 17, 19, 16, 19, 20, 20, 20, 20, 21, 21, 23, 20, 20, 17, 18, 20, 17, 20, 20, 20]
+            "samples": [20, 16, 20, 19, 38, 19, 20, 21, 16, 20, 21, 17, 19, 16, 19, 20, 20, 20, 20, 21, 21, 23, 20, 20, 17, 18, 20, 17, 20, 20, 20, 21]
           },
           "last_updated": "2026-09-21"
         },
@@ -275,7 +275,7 @@
             "min": 16,
             "max": 22,
             "n": 32,
-            "samples": [20, 20, 18, 20, 19, 22, 17, 19, 19, 19, 20, 20, 16, 17, 17, 20, 20, 18, 19, 18, 17, 18, 16, 20, 20, 19, 21, 20, 20, 20, 20, 19]
+            "samples": [20, 18, 20, 19, 22, 17, 19, 19, 19, 20, 20, 16, 17, 17, 20, 20, 18, 19, 18, 17, 18, 16, 20, 20, 19, 21, 20, 20, 20, 20, 19, 21]
           },
           "last_updated": "2026-09-21"
         },
@@ -370,7 +370,7 @@
             "min": 16,
             "max": 29,
             "n": 32,
-            "samples": [21, 20, 20, 20, 19, 19, 29, 20, 20, 19, 20, 21, 16, 17, 17, 20, 20, 16, 19, 19, 18, 21, 17, 20, 20, 21, 19, 20, 20, 19, 20, 20]
+            "samples": [20, 20, 20, 19, 19, 29, 20, 20, 19, 20, 21, 16, 17, 17, 20, 20, 16, 19, 19, 18, 21, 17, 20, 20, 21, 19, 20, 20, 19, 20, 20, 21]
           },
           "last_updated": "2026-09-21"
         },
@@ -465,7 +465,7 @@
             "min": 16,
             "max": 28,
             "n": 32,
-            "samples": [19, 20, 20, 20, 19, 20, 18, 19, 19, 20, 20, 20, 16, 17, 16, 20, 19, 17, 19, 19, 18, 19, 19, 20, 28, 21, 22, 20, 20, 20, 20, 20]
+            "samples": [20, 20, 20, 19, 20, 18, 19, 19, 20, 20, 20, 16, 17, 16, 20, 19, 17, 19, 19, 18, 19, 19, 20, 28, 21, 22, 20, 20, 20, 20, 20, 21]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Effects_composition.json b/test/scenarios/light/scenario_Effects_composition.json
index 80d5bcf0..94f7423c 100644
--- a/test/scenarios/light/scenario_Effects_composition.json
+++ b/test/scenarios/light/scenario_Effects_composition.json
@@ -111,7 +111,7 @@
             "min": 141,
             "max": 656,
             "n": 32,
-            "samples": [145, 144, 144, 144, 145, 147, 143, 158, 656, 144, 144, 144, 143, 148, 150, 144, 142, 147, 146, 152, 143, 149, 160, 145, 142, 144, 145, 142, 141, 143, 143, 144]
+            "samples": [144, 144, 144, 145, 147, 143, 158, 656, 144, 144, 144, 143, 148, 150, 144, 142, 147, 146, 152, 143, 149, 160, 145, 142, 144, 145, 142, 141, 143, 143, 144, 148]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Fields_polar_lut.json b/test/scenarios/light/scenario_Fields_polar_lut.json
index d49f789e..44dfda57 100644
--- a/test/scenarios/light/scenario_Fields_polar_lut.json
+++ b/test/scenarios/light/scenario_Fields_polar_lut.json
@@ -85,12 +85,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 285,
+            "p50": 286,
             "p95": 369,
             "min": 275,
             "max": 421,
             "n": 32,
-            "samples": [279, 278, 290, 293, 285, 285, 283, 286, 284, 421, 369, 286, 303, 285, 284, 289, 295, 307, 292, 287, 305, 299, 287, 335, 282, 284, 282, 276, 275, 282, 283, 283]
+            "samples": [278, 290, 293, 285, 285, 283, 286, 284, 421, 369, 286, 303, 285, 284, 289, 295, 307, 292, 287, 305, 299, 287, 335, 282, 284, 282, 276, 275, 282, 283, 283, 290]
           },
           "last_updated": "2026-09-21"
         }
@@ -115,12 +115,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 284,
+            "p50": 285,
             "p95": 524,
             "min": 274,
             "max": 596,
             "n": 32,
-            "samples": [276, 276, 290, 290, 283, 284, 287, 283, 284, 308, 524, 284, 306, 281, 596, 303, 296, 299, 291, 289, 318, 300, 282, 284, 281, 283, 281, 274, 294, 282, 285, 284]
+            "samples": [276, 290, 290, 283, 284, 287, 283, 284, 308, 524, 284, 306, 281, 596, 303, 296, 299, 291, 289, 318, 300, 282, 284, 281, 283, 281, 274, 294, 282, 285, 284, 293]
           },
           "last_updated": "2026-09-21"
         }
@@ -145,12 +145,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 284,
+            "p50": 285,
             "p95": 455,
-            "min": 277,
+            "min": 278,
             "max": 732,
             "n": 32,
-            "samples": [277, 281, 286, 283, 284, 283, 289, 280, 282, 297, 372, 732, 284, 282, 313, 295, 376, 319, 292, 291, 455, 317, 279, 307, 281, 282, 281, 278, 301, 285, 283, 283]
+            "samples": [281, 286, 283, 284, 283, 289, 280, 282, 297, 372, 732, 284, 282, 313, 295, 376, 319, 292, 291, 455, 317, 279, 307, 281, 282, 281, 278, 301, 285, 283, 283, 289]
           },
           "last_updated": "2026-09-21"
         }
@@ -180,7 +180,7 @@
             "min": 273,
             "max": 607,
             "n": 32,
-            "samples": [275, 278, 283, 282, 286, 287, 283, 282, 281, 297, 324, 343, 284, 303, 307, 304, 371, 305, 607, 289, 326, 302, 273, 282, 282, 281, 282, 276, 274, 283, 283, 282]
+            "samples": [278, 283, 282, 286, 287, 283, 282, 281, 297, 324, 343, 284, 303, 307, 304, 371, 305, 607, 289, 326, 302, 273, 282, 282, 281, 282, 276, 274, 283, 283, 282, 293]
           },
           "last_updated": "2026-09-21"
         }
@@ -205,12 +205,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 140,
+            "p50": 141,
             "p95": 250,
             "min": 136,
             "max": 408,
             "n": 32,
-            "samples": [139, 136, 141, 139, 140, 141, 141, 139, 139, 146, 156, 142, 140, 149, 197, 147, 149, 159, 408, 250, 141, 145, 143, 140, 140, 140, 139, 136, 136, 140, 140, 139]
+            "samples": [136, 141, 139, 140, 141, 141, 139, 139, 146, 156, 142, 140, 149, 197, 147, 149, 159, 408, 250, 141, 145, 143, 140, 140, 140, 139, 136, 136, 140, 140, 139, 148]
           },
           "last_updated": "2026-09-21"
         }
@@ -235,12 +235,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 424,
+            "p50": 434,
             "p95": 562,
             "min": 408,
             "max": 641,
             "n": 32,
-            "samples": [411, 414, 439, 421, 420, 446, 463, 419, 418, 434, 492, 424, 424, 470, 641, 447, 442, 562, 485, 488, 435, 449, 439, 420, 422, 415, 439, 408, 411, 421, 421, 421]
+            "samples": [414, 439, 421, 420, 446, 463, 419, 418, 434, 492, 424, 424, 470, 641, 447, 442, 562, 485, 488, 435, 449, 439, 420, 422, 415, 439, 408, 411, 421, 421, 421, 510]
           },
           "last_updated": "2026-09-21"
         }
@@ -276,7 +276,7 @@
             "min": 1198,
             "max": 2014,
             "n": 32,
-            "samples": [1229, 1219, 1531, 1219, 1451, 1439, 1624, 1275, 1237, 1551, 1340, 1223, 1253, 2014, 1265, 1319, 1459, 1282, 1421, 1365, 1251, 1287, 1262, 1209, 1309, 1263, 1253, 1198, 1218, 1251, 1239, 1224]
+            "samples": [1219, 1531, 1219, 1451, 1439, 1624, 1275, 1237, 1551, 1340, 1223, 1253, 2014, 1265, 1319, 1459, 1282, 1421, 1365, 1251, 1287, 1262, 1209, 1309, 1263, 1253, 1198, 1218, 1251, 1239, 1224, 1243]
           },
           "last_updated": "2026-09-21"
         }
@@ -301,12 +301,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 476,
+            "p50": 477,
             "p95": 698,
             "min": 457,
             "max": 1029,
             "n": 32,
-            "samples": [464, 461, 478, 473, 474, 475, 479, 476, 482, 482, 520, 474, 479, 698, 509, 483, 501, 1029, 473, 532, 472, 475, 488, 468, 525, 473, 471, 477, 457, 473, 471, 491]
+            "samples": [461, 478, 473, 474, 475, 479, 476, 482, 482, 520, 474, 479, 698, 509, 483, 501, 1029, 473, 532, 472, 475, 488, 468, 525, 473, 471, 477, 457, 473, 471, 491, 481]
           },
           "last_updated": "2026-09-21"
         }
diff --git a/test/scenarios/light/scenario_Fluid_solver.json b/test/scenarios/light/scenario_Fluid_solver.json
index 10826896..38c83e71 100644
--- a/test/scenarios/light/scenario_Fluid_solver.json
+++ b/test/scenarios/light/scenario_Fluid_solver.json
@@ -83,7 +83,7 @@
             "min": 28,
             "max": 76,
             "n": 32,
-            "samples": [28, 28, 29, 29, 29, 30, 29, 30, 29, 33, 31, 29, 29, 76, 29, 29, 30, 30, 30, 35, 30, 30, 29, 29, 33, 29, 29, 32, 28, 29, 29, 30]
+            "samples": [28, 29, 29, 29, 30, 29, 30, 29, 33, 31, 29, 29, 76, 29, 29, 30, 30, 30, 35, 30, 30, 29, 29, 33, 29, 29, 32, 28, 29, 29, 30, 29]
           },
           "last_updated": "2026-09-21"
         }
@@ -105,7 +105,7 @@
             "min": 18,
             "max": 196,
             "n": 32,
-            "samples": [18, 19, 19, 19, 19, 19, 19, 19, 19, 21, 21, 19, 19, 196, 19, 19, 20, 23, 19, 24, 19, 19, 19, 19, 21, 19, 19, 32, 18, 19, 19, 19]
+            "samples": [19, 19, 19, 19, 19, 19, 19, 19, 21, 21, 19, 19, 196, 19, 19, 20, 23, 19, 24, 19, 19, 19, 19, 21, 19, 19, 32, 18, 19, 19, 19, 19]
           },
           "last_updated": "2026-09-21"
         }
@@ -127,7 +127,7 @@
             "min": 64,
             "max": 472,
             "n": 32,
-            "samples": [67, 66, 66, 66, 66, 66, 67, 66, 67, 72, 72, 67, 67, 115, 66, 66, 75, 71, 74, 77, 67, 70, 66, 66, 82, 67, 67, 472, 64, 66, 66, 66]
+            "samples": [66, 66, 66, 66, 66, 67, 66, 67, 72, 72, 67, 67, 115, 66, 66, 75, 71, 74, 77, 67, 70, 66, 66, 82, 67, 67, 472, 64, 66, 66, 66, 68]
           },
           "last_updated": "2026-09-21"
         }
@@ -149,7 +149,7 @@
             "min": 28,
             "max": 87,
             "n": 32,
-            "samples": [29, 28, 29, 29, 29, 29, 29, 29, 29, 32, 32, 29, 28, 87, 30, 29, 30, 30, 29, 38, 29, 30, 29, 29, 38, 29, 28, 40, 28, 29, 29, 29]
+            "samples": [28, 29, 29, 29, 29, 29, 29, 29, 32, 32, 29, 28, 87, 30, 29, 30, 30, 29, 38, 29, 30, 29, 29, 38, 29, 28, 40, 28, 29, 29, 29, 29]
           },
           "last_updated": "2026-09-21"
         }
@@ -171,7 +171,7 @@
             "min": 61,
             "max": 118,
             "n": 32,
-            "samples": [61, 61, 64, 62, 62, 63, 63, 62, 64, 65, 70, 62, 63, 118, 63, 62, 67, 66, 67, 72, 63, 69, 62, 63, 80, 62, 62, 74, 62, 64, 63, 63]
+            "samples": [61, 64, 62, 62, 63, 63, 62, 64, 65, 70, 62, 63, 118, 63, 62, 67, 66, 67, 72, 63, 69, 62, 63, 80, 62, 62, 74, 62, 64, 63, 63, 64]
           },
           "last_updated": "2026-09-21"
         }
@@ -192,7 +192,7 @@
             "min": 125,
             "max": 175,
             "n": 32,
-            "samples": [126, 127, 129, 128, 130, 129, 129, 127, 166, 131, 146, 127, 129, 160, 127, 127, 134, 134, 134, 175, 128, 135, 127, 129, 159, 128, 127, 153, 125, 127, 128, 127]
+            "samples": [127, 129, 128, 130, 129, 129, 127, 166, 131, 146, 127, 129, 160, 127, 127, 134, 134, 134, 175, 128, 135, 127, 129, 159, 128, 127, 153, 125, 127, 128, 127, 130]
           },
           "last_updated": "2026-09-21"
         }
@@ -215,7 +215,7 @@
             "min": 127,
             "max": 157,
             "n": 32,
-            "samples": [129, 127, 129, 129, 129, 129, 130, 128, 130, 133, 157, 129, 129, 149, 129, 129, 136, 134, 129, 136, 129, 134, 128, 128, 141, 128, 128, 155, 127, 128, 129, 128]
+            "samples": [127, 129, 129, 129, 129, 130, 128, 130, 133, 157, 129, 129, 149, 129, 129, 136, 134, 129, 136, 129, 134, 128, 128, 141, 128, 128, 155, 127, 128, 129, 128, 131]
           },
           "last_updated": "2026-09-21"
         }
@@ -234,10 +234,10 @@
           "tick_us": {
             "p50": 129,
             "p95": 184,
-            "min": 125,
+            "min": 127,
             "max": 472,
             "n": 32,
-            "samples": [125, 128, 129, 127, 130, 129, 472, 127, 130, 131, 143, 129, 128, 184, 128, 127, 133, 136, 128, 136, 131, 136, 128, 127, 138, 132, 127, 177, 127, 127, 128, 129]
+            "samples": [128, 129, 127, 130, 129, 472, 127, 130, 131, 143, 129, 128, 184, 128, 127, 133, 136, 128, 136, 131, 136, 128, 127, 138, 132, 127, 177, 127, 127, 128, 129, 131]
           },
           "last_updated": "2026-09-21"
         }
@@ -259,7 +259,7 @@
             "min": 34,
             "max": 47,
             "n": 32,
-            "samples": [35, 35, 35, 35, 35, 35, 34, 34, 37, 35, 37, 34, 35, 47, 35, 34, 36, 36, 35, 37, 35, 38, 34, 35, 38, 39, 34, 37, 34, 34, 35, 35]
+            "samples": [35, 35, 35, 35, 35, 34, 34, 37, 35, 37, 34, 35, 47, 35, 34, 36, 36, 35, 37, 35, 38, 34, 35, 38, 39, 34, 37, 34, 34, 35, 35, 36]
           },
           "last_updated": "2026-09-21"
         }
@@ -280,7 +280,7 @@
             "min": 10,
             "max": 26,
             "n": 32,
-            "samples": [10, 10, 10, 10, 10, 10, 11, 10, 11, 11, 11, 11, 10, 26, 10, 10, 11, 11, 11, 11, 10, 11, 10, 10, 14, 10, 10, 11, 10, 10, 10, 10]
+            "samples": [10, 10, 10, 10, 10, 11, 10, 11, 11, 11, 11, 10, 26, 10, 10, 11, 11, 11, 11, 10, 11, 10, 10, 14, 10, 10, 11, 10, 10, 10, 10, 11]
           },
           "last_updated": "2026-09-21"
         }
@@ -301,7 +301,7 @@
             "min": 214,
             "max": 335,
             "n": 32,
-            "samples": [214, 218, 217, 220, 216, 216, 217, 215, 231, 221, 235, 215, 245, 335, 219, 214, 224, 231, 218, 231, 218, 225, 216, 216, 250, 215, 214, 216, 214, 216, 216, 216]
+            "samples": [218, 217, 220, 216, 216, 217, 215, 231, 221, 235, 215, 245, 335, 219, 214, 224, 231, 218, 231, 218, 225, 216, 216, 250, 215, 214, 216, 214, 216, 216, 216, 220]
           },
           "last_updated": "2026-09-21"
         }
@@ -323,7 +323,7 @@
             "min": 10,
             "max": 12,
             "n": 32,
-            "samples": [11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 11, 11, 10, 11, 11, 11, 11, 12, 10, 11, 10, 10, 11, 10, 11, 10, 11, 11, 11, 10]
+            "samples": [11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 11, 11, 10, 11, 11, 11, 11, 12, 10, 11, 10, 10, 11, 10, 11, 10, 11, 11, 11, 10, 11]
           },
           "last_updated": "2026-09-21"
         }
@@ -345,7 +345,7 @@
             "min": 8,
             "max": 9,
             "n": 32,
-            "samples": [9, 8, 8, 9, 8, 8, 8, 8, 9, 9, 9, 8, 9, 9, 8, 8, 9, 9, 8, 9, 8, 8, 8, 8, 9, 8, 8, 9, 8, 8, 8, 9]
+            "samples": [8, 8, 9, 8, 8, 8, 8, 9, 9, 9, 8, 9, 9, 8, 8, 9, 9, 8, 9, 8, 8, 8, 8, 9, 8, 8, 9, 8, 8, 8, 9, 9]
           },
           "last_updated": "2026-09-21"
         }
@@ -366,7 +366,7 @@
             "min": 6,
             "max": 8,
             "n": 32,
-            "samples": [7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 7, 7, 7, 7, 7, 7, 7, 7, 8, 7, 6, 7, 7, 7, 6, 7]
+            "samples": [7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 7, 7, 7, 7, 7, 7, 7, 7, 8, 7, 6, 7, 7, 7, 6, 7, 7]
           },
           "last_updated": "2026-09-21"
         }
diff --git a/test/scenarios/light/scenario_GridLayout_resize.json b/test/scenarios/light/scenario_GridLayout_resize.json
index 51fe07d7..2aa4db40 100644
--- a/test/scenarios/light/scenario_GridLayout_resize.json
+++ b/test/scenarios/light/scenario_GridLayout_resize.json
@@ -122,7 +122,7 @@
             "min": 117,
             "max": 158,
             "n": 32,
-            "samples": [117, 120, 121, 120, 118, 123, 124, 120, 124, 126, 132, 120, 142, 128, 119, 118, 124, 118, 119, 129, 121, 126, 117, 119, 158, 122, 120, 117, 118, 120, 119, 119]
+            "samples": [120, 121, 120, 118, 123, 124, 120, 124, 126, 132, 120, 142, 128, 119, 118, 124, 118, 119, 129, 121, 126, 117, 119, 158, 122, 120, 117, 118, 120, 119, 119, 123]
           },
           "last_updated": "2026-09-21"
         },
@@ -304,7 +304,7 @@
             "min": 59,
             "max": 86,
             "n": 32,
-            "samples": [64, 68, 65, 61, 65, 65, 66, 64, 68, 66, 64, 64, 76, 86, 74, 65, 60, 60, 64, 66, 64, 61, 64, 65, 65, 64, 64, 59, 64, 64, 64, 65]
+            "samples": [68, 65, 61, 65, 65, 66, 64, 68, 66, 64, 64, 76, 86, 74, 65, 60, 60, 64, 66, 64, 61, 64, 65, 65, 64, 64, 59, 64, 64, 64, 65, 67]
           },
           "last_updated": "2026-09-21"
         },
@@ -486,7 +486,7 @@
             "min": 116,
             "max": 263,
             "n": 32,
-            "samples": [118, 120, 120, 124, 119, 122, 121, 120, 125, 121, 120, 119, 128, 263, 119, 119, 126, 118, 120, 126, 122, 124, 120, 120, 123, 119, 120, 116, 119, 120, 119, 119]
+            "samples": [120, 120, 124, 119, 122, 121, 120, 125, 121, 120, 119, 128, 263, 119, 119, 126, 118, 120, 126, 122, 124, 120, 120, 123, 119, 120, 116, 119, 120, 119, 119, 123]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Layer_base_pipeline.json b/test/scenarios/light/scenario_Layer_base_pipeline.json
index bb2ced5d..1508acc5 100644
--- a/test/scenarios/light/scenario_Layer_base_pipeline.json
+++ b/test/scenarios/light/scenario_Layer_base_pipeline.json
@@ -88,7 +88,7 @@
             "min": 63,
             "max": 87,
             "n": 32,
-            "samples": [70, 71, 67, 71, 68, 70, 66, 71, 70, 70, 70, 66, 87, 70, 69, 71, 67, 70, 74, 70, 69, 69, 71, 72, 70, 69, 64, 63, 67, 71, 68, 69]
+            "samples": [71, 67, 71, 68, 70, 66, 71, 70, 70, 70, 66, 87, 70, 69, 71, 67, 70, 74, 70, 69, 69, 71, 72, 70, 69, 64, 63, 67, 71, 68, 69, 72]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Layer_memory_1to1.json b/test/scenarios/light/scenario_Layer_memory_1to1.json
index 25cafdb0..4d54c4f5 100644
--- a/test/scenarios/light/scenario_Layer_memory_1to1.json
+++ b/test/scenarios/light/scenario_Layer_memory_1to1.json
@@ -85,7 +85,7 @@
             "min": 4,
             "max": 28,
             "n": 32,
-            "samples": [4, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 24, 8, 5, 5, 5, 28, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 24, 8, 5, 5, 5, 28, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Layouts_mutation.json b/test/scenarios/light/scenario_Layouts_mutation.json
index e8f1c502..ec328863 100644
--- a/test/scenarios/light/scenario_Layouts_mutation.json
+++ b/test/scenarios/light/scenario_Layouts_mutation.json
@@ -83,7 +83,7 @@
             "min": 16,
             "max": 45,
             "n": 32,
-            "samples": [17, 17, 17, 18, 16, 16, 16, 17, 19, 17, 16, 45, 37, 16, 16, 17, 20, 16, 17, 16, 18, 16, 16, 17, 16, 16, 17, 16, 16, 16, 16, 16]
+            "samples": [17, 17, 18, 16, 16, 16, 17, 19, 17, 16, 45, 37, 16, 16, 17, 20, 16, 17, 16, 18, 16, 16, 17, 16, 16, 17, 16, 16, 16, 16, 16, 18]
           },
           "last_updated": "2026-09-21"
         },
@@ -211,7 +211,7 @@
             "min": 44,
             "max": 77,
             "n": 32,
-            "samples": [50, 50, 58, 51, 49, 47, 48, 49, 53, 56, 49, 69, 52, 49, 47, 51, 77, 50, 49, 50, 50, 47, 50, 51, 49, 50, 44, 49, 46, 49, 46, 49]
+            "samples": [50, 58, 51, 49, 47, 48, 49, 53, 56, 49, 69, 52, 49, 47, 51, 77, 50, 49, 50, 50, 47, 50, 51, 49, 50, 44, 49, 46, 49, 46, 49, 53]
           },
           "last_updated": "2026-09-21"
         },
@@ -334,7 +334,7 @@
             "min": 90,
             "max": 104,
             "n": 32,
-            "samples": [94, 94, 92, 97, 91, 93, 92, 93, 95, 99, 93, 104, 99, 92, 92, 90, 90, 93, 100, 93, 93, 92, 93, 96, 93, 93, 91, 92, 93, 92, 93, 94]
+            "samples": [94, 92, 97, 91, 93, 92, 93, 95, 99, 93, 104, 99, 92, 92, 90, 90, 93, 100, 93, 93, 92, 93, 96, 93, 93, 91, 92, 93, 92, 93, 94, 96]
           },
           "last_updated": "2026-09-21"
         },
@@ -456,7 +456,7 @@
             "min": 16,
             "max": 23,
             "n": 32,
-            "samples": [20, 19, 19, 21, 20, 20, 19, 20, 20, 17, 20, 17, 20, 19, 20, 23, 16, 19, 17, 20, 16, 21, 20, 20, 20, 20, 17, 16, 18, 19, 20, 20]
+            "samples": [19, 19, 21, 20, 20, 19, 20, 20, 17, 20, 17, 20, 19, 20, 23, 16, 19, 17, 20, 16, 21, 20, 20, 20, 20, 17, 16, 18, 19, 20, 20, 21]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_MoonLiveEffect_livescript.json b/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
index ebdd2a5e..3f7866f6 100644
--- a/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
+++ b/test/scenarios/light/scenario_MoonLiveEffect_livescript.json
@@ -93,7 +93,7 @@
             "min": 5,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 6, 5]
+            "samples": [5, 5, 5, 5, 5, 6, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 6, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -211,7 +211,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 4, 5, 6, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 4, 5, 6, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -321,7 +321,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 4, 5, 6, 6, 6, 5, 5, 5, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 6, 6, 6, 5, 5, 5, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -431,7 +431,7 @@
             "min": 5,
             "max": 7,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 5, 5, 6, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 6, 6, 6, 5, 5, 6, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -534,7 +534,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 4, 4, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 4, 4, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -637,7 +637,7 @@
             "min": 4,
             "max": 10,
             "n": 32,
-            "samples": [5, 5, 5, 5, 6, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 4, 6, 7, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5]
+            "samples": [5, 5, 5, 6, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 4, 6, 7, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -740,7 +740,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 4, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 4, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         }
@@ -762,7 +762,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 6, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         }
@@ -784,7 +784,7 @@
             "min": 4,
             "max": 8,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 6, 5, 5, 6, 5, 5, 5, 8, 5, 5, 5, 6]
+            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 6, 5, 5, 6, 5, 5, 5, 8, 5, 5, 5, 6, 5]
           },
           "last_updated": "2026-09-21"
         }
@@ -804,7 +804,7 @@
             "min": 5,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 6, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -907,7 +907,7 @@
             "min": 4,
             "max": 9,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 4, 5, 4, 7, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 4, 5, 4, 7, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -1010,7 +1010,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_MoonLive_pipeline.json b/test/scenarios/light/scenario_MoonLive_pipeline.json
index 2f2a1fa2..2747373e 100644
--- a/test/scenarios/light/scenario_MoonLive_pipeline.json
+++ b/test/scenarios/light/scenario_MoonLive_pipeline.json
@@ -379,7 +379,7 @@
             "min": 4,
             "max": 9,
             "n": 32,
-            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 4, 5]
+            "samples": [5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 9, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -536,7 +536,7 @@
             "min": 4,
             "max": 11,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 11, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 11, 5, 5, 5, 4, 6, 5, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 6, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -687,7 +687,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 6, 6, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -989,7 +989,7 @@
             "min": 5,
             "max": 10,
             "n": 32,
-            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5]
+            "samples": [5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 10, 5, 5, 5, 5, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
@@ -1131,7 +1131,7 @@
             "min": 4,
             "max": 7,
             "n": 32,
-            "samples": [5, 5, 5, 5, 4, 5, 5, 5, 4, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5]
+            "samples": [5, 5, 5, 4, 5, 5, 5, 4, 5, 5, 5, 7, 5, 5, 5, 5, 5, 5, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 5, 5, 5]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json b/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
index e59fd3e7..beac1cdf 100644
--- a/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
+++ b/test/scenarios/light/scenario_MultiplyModifier_memory_lut.json
@@ -94,7 +94,7 @@
             "min": 2,
             "max": 4,
             "n": 32,
-            "samples": [3, 3, 4, 3, 3, 3, 3, 3, 2, 3, 3, 3, 3, 3, 2, 2, 3, 3, 3, 3, 2, 3, 3, 3, 3, 2, 3, 3, 2, 3, 3, 3]
+            "samples": [3, 4, 3, 3, 3, 3, 3, 2, 3, 3, 3, 3, 3, 2, 2, 3, 3, 3, 3, 2, 3, 3, 3, 3, 2, 3, 3, 2, 3, 3, 3, 3]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_MultiplyModifier_pipeline.json b/test/scenarios/light/scenario_MultiplyModifier_pipeline.json
index 8f75331b..4f97b94d 100644
--- a/test/scenarios/light/scenario_MultiplyModifier_pipeline.json
+++ b/test/scenarios/light/scenario_MultiplyModifier_pipeline.json
@@ -94,7 +94,7 @@
             "min": 117,
             "max": 158,
             "n": 32,
-            "samples": [118, 117, 120, 119, 121, 119, 119, 118, 118, 119, 121, 118, 158, 131, 117, 117, 120, 123, 121, 142, 118, 127, 118, 118, 122, 122, 117, 117, 117, 119, 117, 118]
+            "samples": [117, 120, 119, 121, 119, 119, 118, 118, 119, 121, 118, 158, 131, 117, 117, 120, 123, 121, 142, 118, 127, 118, 118, 122, 122, 117, 117, 117, 119, 117, 118, 121]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_Trails_ladder.json b/test/scenarios/light/scenario_Trails_ladder.json
index 671f9f20..ee551f5c 100644
--- a/test/scenarios/light/scenario_Trails_ladder.json
+++ b/test/scenarios/light/scenario_Trails_ladder.json
@@ -91,7 +91,7 @@
             "min": 11,
             "max": 13,
             "n": 32,
-            "samples": [11, 11, 12, 11, 11, 11, 11, 11, 11, 11, 13, 12, 12, 11, 11, 11, 11, 13, 11, 11, 11, 12, 11, 11, 12, 11, 11, 11, 11, 11, 11, 11]
+            "samples": [11, 12, 11, 11, 11, 11, 11, 11, 11, 13, 12, 12, 11, 11, 11, 11, 13, 11, 11, 11, 12, 11, 11, 12, 11, 11, 11, 11, 11, 11, 11, 12]
           },
           "last_updated": "2026-09-21"
         }
@@ -128,7 +128,7 @@
             "min": 44,
             "max": 55,
             "n": 32,
-            "samples": [44, 44, 46, 45, 44, 46, 47, 44, 45, 46, 47, 47, 46, 47, 45, 44, 46, 47, 44, 46, 44, 46, 45, 45, 55, 46, 44, 45, 44, 44, 45, 45]
+            "samples": [44, 46, 45, 44, 46, 47, 44, 45, 46, 47, 47, 46, 47, 45, 44, 46, 47, 44, 46, 44, 46, 45, 45, 55, 46, 44, 45, 44, 44, 45, 45, 46]
           },
           "last_updated": "2026-09-21"
         }
@@ -165,7 +165,7 @@
             "min": 171,
             "max": 274,
             "n": 32,
-            "samples": [177, 175, 178, 177, 179, 178, 178, 176, 177, 189, 189, 187, 180, 184, 176, 176, 185, 191, 178, 189, 178, 191, 176, 176, 274, 188, 178, 198, 171, 176, 176, 178]
+            "samples": [175, 178, 177, 179, 178, 178, 176, 177, 189, 189, 187, 180, 184, 176, 176, 185, 191, 178, 189, 178, 191, 176, 176, 274, 188, 178, 198, 171, 176, 176, 178, 181]
           },
           "last_updated": "2026-09-21"
         }
@@ -204,12 +204,12 @@
       "observed": {
         "desktop-macos": {
           "tick_us": {
-            "p50": 347,
+            "p50": 348,
             "p95": 421,
             "min": 335,
             "max": 466,
             "n": 32,
-            "samples": [345, 346, 356, 347, 347, 348, 347, 466, 347, 351, 358, 348, 399, 362, 348, 346, 358, 365, 346, 361, 345, 359, 346, 347, 421, 355, 346, 362, 335, 347, 347, 347]
+            "samples": [346, 356, 347, 347, 348, 347, 466, 347, 351, 358, 348, 399, 362, 348, 346, 358, 365, 346, 361, 345, 359, 346, 347, 421, 355, 346, 362, 335, 347, 347, 347, 357]
           },
           "last_updated": "2026-09-21"
         }
@@ -239,7 +239,7 @@
             "min": 348,
             "max": 742,
             "n": 32,
-            "samples": [358, 352, 359, 353, 354, 356, 356, 352, 352, 359, 523, 376, 365, 368, 658, 362, 367, 371, 354, 370, 354, 369, 355, 358, 488, 359, 354, 742, 348, 355, 355, 357]
+            "samples": [352, 359, 353, 354, 356, 356, 352, 352, 359, 523, 376, 365, 368, 658, 362, 367, 371, 354, 370, 354, 369, 355, 358, 488, 359, 354, 742, 348, 355, 355, 357, 364]
           },
           "last_updated": "2026-09-21"
         }
@@ -269,7 +269,7 @@
             "min": 345,
             "max": 426,
             "n": 32,
-            "samples": [352, 347, 350, 346, 348, 400, 349, 346, 346, 351, 426, 347, 357, 349, 350, 354, 364, 360, 346, 358, 346, 370, 351, 349, 393, 352, 345, 349, 345, 347, 347, 360]
+            "samples": [347, 350, 346, 348, 400, 349, 346, 346, 351, 426, 347, 357, 349, 350, 354, 364, 360, 346, 358, 346, 370, 351, 349, 393, 352, 345, 349, 345, 347, 347, 360, 353]
           },
           "last_updated": "2026-09-21"
         }
@@ -313,7 +313,7 @@
             "min": 176,
             "max": 268,
             "n": 32,
-            "samples": [178, 177, 177, 177, 179, 268, 178, 176, 176, 179, 186, 176, 178, 177, 182, 176, 184, 179, 177, 183, 177, 185, 176, 177, 177, 178, 176, 176, 176, 177, 176, 184]
+            "samples": [177, 177, 177, 179, 268, 178, 176, 176, 179, 186, 176, 178, 177, 182, 176, 184, 179, 177, 183, 177, 185, 176, 177, 177, 178, 176, 176, 176, 177, 176, 184, 180]
           },
           "last_updated": "2026-09-21"
         }
diff --git a/test/scenarios/light/scenario_modifier_chain.json b/test/scenarios/light/scenario_modifier_chain.json
index c4c39498..09a74331 100644
--- a/test/scenarios/light/scenario_modifier_chain.json
+++ b/test/scenarios/light/scenario_modifier_chain.json
@@ -107,7 +107,7 @@
             "min": 8,
             "max": 27,
             "n": 32,
-            "samples": [8, 8, 10, 10, 9, 10, 8, 10, 9, 11, 9, 10, 10, 10, 9, 10, 8, 10, 27, 9, 10, 10, 8, 10, 10, 10, 8, 8, 10, 10, 10, 10]
+            "samples": [8, 10, 10, 9, 10, 8, 10, 9, 11, 9, 10, 10, 10, 9, 10, 8, 10, 27, 9, 10, 10, 8, 10, 10, 10, 8, 8, 10, 10, 10, 10, 10]
           },
           "last_updated": "2026-09-21"
         },
@@ -167,7 +167,7 @@
             "min": 6,
             "max": 11,
             "n": 32,
-            "samples": [9, 7, 9, 7, 9, 9, 6, 9, 9, 9, 9, 9, 9, 9, 8, 9, 7, 9, 11, 9, 9, 9, 9, 8, 9, 10, 7, 9, 8, 9, 9, 10]
+            "samples": [7, 9, 7, 9, 9, 6, 9, 9, 9, 9, 9, 9, 9, 8, 9, 7, 9, 11, 9, 9, 9, 9, 8, 9, 10, 7, 9, 8, 9, 9, 10, 9]
           },
           "last_updated": "2026-09-21"
         },
@@ -225,7 +225,7 @@
             "min": 21,
             "max": 25,
             "n": 32,
-            "samples": [24, 22, 24, 21, 25, 25, 21, 24, 25, 24, 24, 24, 24, 24, 24, 24, 21, 24, 24, 25, 25, 25, 24, 24, 24, 25, 21, 23, 24, 24, 24, 24]
+            "samples": [22, 24, 21, 25, 25, 21, 24, 25, 24, 24, 24, 24, 24, 24, 24, 21, 24, 24, 25, 25, 25, 24, 24, 24, 25, 21, 23, 24, 24, 24, 24, 25]
           },
           "last_updated": "2026-09-21"
         },
@@ -254,11 +254,11 @@
         "desktop-macos": {
           "tick_us": {
             "p50": 43,
-            "p95": 46,
+            "p95": 47,
             "min": 35,
             "max": 47,
             "n": 32,
-            "samples": [44, 38, 44, 47, 42, 43, 45, 43, 45, 44, 43, 43, 43, 42, 44, 45, 37, 44, 44, 41, 43, 43, 44, 39, 45, 45, 36, 35, 46, 44, 42, 44]
+            "samples": [38, 44, 47, 42, 43, 45, 43, 45, 44, 43, 43, 43, 42, 44, 45, 37, 44, 44, 41, 43, 43, 44, 39, 45, 45, 36, 35, 46, 44, 42, 44, 47]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_modifier_swap.json b/test/scenarios/light/scenario_modifier_swap.json
index 6d4cffe3..daaa3720 100644
--- a/test/scenarios/light/scenario_modifier_swap.json
+++ b/test/scenarios/light/scenario_modifier_swap.json
@@ -156,7 +156,7 @@
             "min": 8,
             "max": 11,
             "n": 32,
-            "samples": [8, 9, 8, 8, 11, 8, 8, 8, 10, 8, 8, 10, 8, 8, 8, 10, 8, 8, 8, 8, 8, 8, 8, 8, 10, 11, 8, 8, 9, 8, 8, 8]
+            "samples": [9, 8, 8, 11, 8, 8, 8, 10, 8, 8, 10, 8, 8, 8, 10, 8, 8, 8, 8, 8, 8, 8, 8, 10, 11, 8, 8, 9, 8, 8, 8, 8]
           },
           "last_updated": "2026-09-21"
         },
@@ -300,7 +300,7 @@
             "min": 20,
             "max": 53,
             "n": 32,
-            "samples": [23, 24, 21, 21, 24, 21, 21, 24, 24, 25, 24, 24, 24, 21, 23, 25, 20, 23, 22, 22, 23, 24, 21, 53, 24, 25, 21, 21, 24, 24, 22, 20]
+            "samples": [24, 21, 21, 24, 21, 21, 24, 24, 25, 24, 24, 24, 21, 23, 25, 20, 23, 22, 22, 23, 24, 21, 53, 24, 25, 21, 21, 24, 24, 22, 20, 22]
           },
           "last_updated": "2026-09-21"
         },
@@ -444,7 +444,7 @@
             "min": 8,
             "max": 16,
             "n": 32,
-            "samples": [10, 10, 10, 8, 8, 10, 10, 10, 10, 11, 10, 10, 10, 8, 10, 10, 9, 10, 11, 10, 11, 10, 10, 16, 10, 10, 8, 11, 10, 10, 10, 10]
+            "samples": [10, 10, 8, 8, 10, 10, 10, 10, 11, 10, 10, 10, 8, 10, 10, 9, 10, 11, 10, 11, 10, 10, 16, 10, 10, 8, 11, 10, 10, 10, 10, 10]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_perf_full.json b/test/scenarios/light/scenario_perf_full.json
index d8784adc..a0a104b9 100644
--- a/test/scenarios/light/scenario_perf_full.json
+++ b/test/scenarios/light/scenario_perf_full.json
@@ -90,7 +90,7 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
           "last_updated": "2026-09-21"
         },
@@ -210,7 +210,7 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
           "last_updated": "2026-09-21"
         },
@@ -330,7 +330,7 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
           "last_updated": "2026-09-21"
         },
@@ -573,7 +573,7 @@
             "min": 1,
             "max": 3,
             "n": 32,
-            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+            "samples": [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
           },
           "last_updated": "2026-09-21"
         },
@@ -1297,7 +1297,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -1421,7 +1421,7 @@
             "min": 17,
             "max": 79,
             "n": 32,
-            "samples": [18, 18, 79, 17, 18, 18, 19, 18, 18, 19, 18, 17, 18, 19, 17, 18, 17, 18, 18, 19, 19, 17, 18, 18, 18, 18, 18, 17, 17, 17, 17, 18]
+            "samples": [18, 79, 17, 18, 18, 19, 18, 18, 19, 18, 17, 18, 19, 17, 18, 17, 18, 18, 19, 19, 17, 18, 18, 18, 18, 18, 17, 17, 17, 17, 18, 18]
           },
           "last_updated": "2026-09-21"
         },
@@ -1545,7 +1545,7 @@
             "min": 70,
             "max": 80,
             "n": 32,
-            "samples": [71, 70, 80, 70, 72, 72, 77, 71, 72, 72, 75, 70, 72, 73, 70, 71, 70, 75, 71, 78, 76, 70, 73, 71, 78, 76, 73, 70, 72, 70, 70, 70]
+            "samples": [70, 80, 70, 72, 72, 77, 71, 72, 72, 75, 70, 72, 73, 70, 71, 70, 75, 71, 78, 76, 70, 73, 71, 78, 76, 73, 70, 72, 70, 70, 70, 72]
           },
           "last_updated": "2026-09-21"
         },
@@ -1677,7 +1677,7 @@
             "min": 4,
             "max": 5,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -1801,7 +1801,7 @@
             "min": 15,
             "max": 18,
             "n": 32,
-            "samples": [16, 15, 16, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 16, 15, 17, 15, 18, 16, 16, 16, 16, 17, 15, 16, 15, 16, 16, 15, 16]
+            "samples": [15, 16, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 16, 15, 17, 15, 18, 16, 16, 16, 16, 17, 15, 16, 15, 16, 16, 15, 16, 16]
           },
           "last_updated": "2026-09-21"
         },
@@ -1925,7 +1925,7 @@
             "min": 60,
             "max": 68,
             "n": 32,
-            "samples": [63, 63, 63, 62, 62, 67, 67, 64, 62, 65, 67, 62, 64, 64, 63, 63, 62, 67, 62, 68, 67, 64, 65, 63, 68, 63, 65, 61, 60, 63, 62, 62]
+            "samples": [63, 63, 62, 62, 67, 67, 64, 62, 65, 67, 62, 64, 64, 63, 63, 62, 67, 62, 68, 67, 64, 65, 63, 68, 63, 65, 61, 60, 63, 62, 62, 64]
           },
           "last_updated": "2026-09-21"
         },
@@ -2049,7 +2049,7 @@
             "min": 243,
             "max": 297,
             "n": 32,
-            "samples": [252, 251, 251, 250, 252, 265, 297, 250, 250, 266, 295, 249, 257, 261, 251, 251, 251, 265, 251, 269, 272, 252, 262, 251, 263, 248, 265, 251, 243, 250, 250, 249]
+            "samples": [251, 251, 250, 252, 265, 297, 250, 250, 266, 295, 249, 257, 261, 251, 251, 251, 265, 251, 269, 272, 252, 262, 251, 263, 248, 265, 251, 243, 250, 250, 249, 256]
           },
           "last_updated": "2026-09-21"
         },
@@ -2456,7 +2456,7 @@
             "min": 15,
             "max": 22,
             "n": 32,
-            "samples": [16, 17, 15, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 15, 15, 16, 15, 16, 22, 17, 17, 15, 15, 15, 16, 17, 15, 15, 15, 15]
+            "samples": [17, 15, 15, 15, 17, 18, 15, 15, 16, 17, 15, 16, 16, 15, 15, 15, 16, 15, 16, 22, 17, 17, 15, 15, 15, 16, 17, 15, 15, 15, 15, 16]
           },
           "last_updated": "2026-09-21"
         },
@@ -2580,7 +2580,7 @@
             "min": 60,
             "max": 82,
             "n": 32,
-            "samples": [63, 62, 62, 62, 63, 68, 71, 62, 62, 82, 75, 62, 64, 66, 62, 63, 62, 65, 64, 63, 70, 63, 65, 62, 62, 62, 65, 64, 60, 62, 62, 63]
+            "samples": [62, 62, 62, 63, 68, 71, 62, 62, 82, 75, 62, 64, 66, 62, 63, 62, 65, 64, 63, 70, 63, 65, 62, 62, 62, 65, 64, 60, 62, 62, 63, 64]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_perf_light.json b/test/scenarios/light/scenario_perf_light.json
index 5b7c7d5c..0c87d9a1 100644
--- a/test/scenarios/light/scenario_perf_light.json
+++ b/test/scenarios/light/scenario_perf_light.json
@@ -578,7 +578,7 @@
             "min": 4,
             "max": 6,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -702,7 +702,7 @@
             "min": 15,
             "max": 27,
             "n": 32,
-            "samples": [15, 15, 15, 15, 15, 15, 16, 17, 27, 15, 15, 16, 18, 15, 16, 16, 15, 16, 15, 16, 17, 16, 16, 15, 15, 21, 17, 16, 15, 15, 15, 15]
+            "samples": [15, 15, 15, 15, 15, 16, 17, 27, 15, 15, 16, 18, 15, 16, 16, 15, 16, 15, 16, 17, 16, 16, 15, 15, 21, 17, 16, 15, 15, 15, 15, 16]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_peripheral_grid_sweep.json b/test/scenarios/light/scenario_peripheral_grid_sweep.json
index 20766163..f5e2ea2a 100644
--- a/test/scenarios/light/scenario_peripheral_grid_sweep.json
+++ b/test/scenarios/light/scenario_peripheral_grid_sweep.json
@@ -178,7 +178,7 @@
             "min": 4,
             "max": 15,
             "n": 32,
-            "samples": [4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -305,7 +305,7 @@
             "min": 15,
             "max": 122,
             "n": 32,
-            "samples": [15, 15, 16, 122, 17, 15, 16, 15, 18, 16, 16, 16, 16, 16, 16, 16, 17, 16, 19, 17, 16, 17, 16, 16, 16, 16, 16, 16, 15, 16, 16, 16]
+            "samples": [15, 16, 122, 17, 15, 16, 15, 18, 16, 16, 16, 16, 16, 16, 16, 17, 16, 19, 17, 16, 17, 16, 16, 16, 16, 16, 16, 15, 16, 16, 16, 16]
           },
           "last_updated": "2026-09-21"
         },
@@ -432,7 +432,7 @@
             "min": 61,
             "max": 72,
             "n": 32,
-            "samples": [63, 62, 64, 71, 69, 63, 62, 64, 67, 62, 64, 67, 63, 63, 63, 64, 70, 68, 72, 69, 66, 67, 63, 65, 63, 62, 69, 65, 61, 62, 69, 63]
+            "samples": [62, 64, 71, 69, 63, 62, 64, 67, 62, 64, 67, 63, 63, 63, 64, 70, 68, 72, 69, 66, 67, 63, 65, 63, 62, 69, 65, 61, 62, 69, 63, 63]
           },
           "last_updated": "2026-09-21"
         },
@@ -559,7 +559,7 @@
             "min": 242,
             "max": 325,
             "n": 32,
-            "samples": [252, 250, 249, 286, 278, 249, 250, 252, 253, 249, 256, 268, 250, 254, 250, 263, 252, 273, 291, 275, 325, 317, 249, 260, 248, 249, 260, 260, 242, 251, 251, 250]
+            "samples": [250, 249, 286, 278, 249, 250, 252, 253, 249, 256, 268, 250, 254, 250, 263, 252, 273, 291, 275, 325, 317, 249, 260, 248, 249, 260, 260, 242, 251, 251, 250, 261]
           },
           "last_updated": "2026-09-21"
         },
@@ -707,7 +707,7 @@
             "min": 4,
             "max": 15,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 15, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -834,7 +834,7 @@
             "min": 15,
             "max": 19,
             "n": 32,
-            "samples": [15, 15, 17, 17, 17, 15, 15, 16, 15, 15, 16, 16, 15, 16, 15, 16, 16, 16, 18, 17, 19, 16, 16, 15, 15, 15, 17, 16, 15, 15, 16, 15]
+            "samples": [15, 17, 17, 17, 15, 15, 16, 15, 15, 16, 16, 15, 16, 15, 16, 16, 16, 18, 17, 19, 16, 16, 15, 15, 15, 17, 16, 15, 15, 16, 15, 16]
           },
           "last_updated": "2026-09-21"
         },
@@ -961,7 +961,7 @@
             "min": 60,
             "max": 81,
             "n": 32,
-            "samples": [62, 62, 62, 67, 67, 62, 62, 62, 63, 62, 64, 67, 63, 63, 62, 68, 64, 66, 73, 67, 81, 68, 62, 62, 61, 63, 65, 63, 60, 62, 63, 62]
+            "samples": [62, 62, 67, 67, 62, 62, 62, 63, 62, 64, 67, 63, 63, 62, 68, 64, 66, 73, 67, 81, 68, 62, 62, 61, 63, 65, 63, 60, 62, 63, 62, 64]
           },
           "last_updated": "2026-09-21"
         },
@@ -1088,7 +1088,7 @@
             "min": 243,
             "max": 286,
             "n": 32,
-            "samples": [252, 249, 249, 268, 267, 252, 251, 252, 251, 248, 259, 269, 251, 252, 250, 263, 250, 277, 285, 270, 286, 264, 249, 253, 249, 253, 255, 250, 243, 250, 251, 250]
+            "samples": [249, 249, 268, 267, 252, 251, 252, 251, 248, 259, 269, 251, 252, 250, 263, 250, 277, 285, 270, 286, 264, 249, 253, 249, 253, 255, 250, 243, 250, 251, 250, 267]
           },
           "last_updated": "2026-09-21"
         },
@@ -1363,7 +1363,7 @@
             "min": 15,
             "max": 17,
             "n": 32,
-            "samples": [16, 15, 16, 17, 16, 15, 15, 16, 15, 16, 16, 16, 15, 15, 15, 16, 15, 17, 17, 16, 17, 16, 16, 16, 15, 15, 16, 15, 15, 15, 15, 15]
+            "samples": [15, 16, 17, 16, 15, 15, 16, 15, 16, 16, 16, 15, 15, 15, 16, 15, 17, 17, 16, 17, 16, 16, 16, 15, 15, 16, 15, 15, 15, 15, 15, 16]
           },
           "last_updated": "2026-09-21"
         },
@@ -1485,12 +1485,12 @@
         },
         "desktop-macos": {
           "tick_us": {
-            "p50": 62,
+            "p50": 63,
             "p95": 69,
             "min": 60,
             "max": 70,
             "n": 32,
-            "samples": [62, 64, 62, 66, 65, 63, 62, 63, 62, 62, 63, 66, 62, 63, 62, 67, 62, 69, 68, 65, 70, 65, 62, 62, 61, 62, 64, 61, 60, 62, 65, 62]
+            "samples": [64, 62, 66, 65, 63, 62, 63, 62, 62, 63, 66, 62, 63, 62, 67, 62, 69, 68, 65, 70, 65, 62, 62, 61, 62, 64, 61, 60, 62, 65, 62, 64]
           },
           "last_updated": "2026-09-21"
         },
@@ -1617,7 +1617,7 @@
             "min": 245,
             "max": 395,
             "n": 32,
-            "samples": [251, 249, 249, 254, 263, 251, 250, 251, 267, 253, 261, 265, 250, 254, 250, 265, 252, 275, 278, 255, 258, 395, 250, 248, 246, 250, 256, 245, 251, 250, 250, 249]
+            "samples": [249, 249, 254, 263, 251, 250, 251, 267, 253, 261, 265, 250, 254, 250, 265, 252, 275, 278, 255, 258, 395, 250, 248, 246, 250, 256, 245, 251, 250, 250, 249, 254]
           },
           "last_updated": "2026-09-21"
         },
@@ -1765,7 +1765,7 @@
             "min": 4,
             "max": 10,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 10, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 6, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 10, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -1887,12 +1887,12 @@
         },
         "desktop-macos": {
           "tick_us": {
-            "p50": 15,
+            "p50": 16,
             "p95": 17,
             "min": 15,
             "max": 32,
             "n": 32,
-            "samples": [15, 15, 15, 16, 16, 15, 15, 15, 17, 15, 17, 16, 16, 15, 16, 16, 15, 17, 17, 16, 16, 17, 15, 15, 16, 15, 16, 15, 32, 15, 15, 15]
+            "samples": [15, 15, 16, 16, 15, 15, 15, 17, 15, 17, 16, 16, 15, 16, 16, 15, 17, 17, 16, 16, 17, 15, 15, 16, 15, 16, 15, 32, 15, 15, 15, 16]
           },
           "last_updated": "2026-09-21"
         },
@@ -2019,7 +2019,7 @@
             "min": 60,
             "max": 114,
             "n": 32,
-            "samples": [62, 62, 62, 63, 68, 62, 64, 62, 68, 62, 65, 65, 62, 63, 62, 64, 62, 69, 67, 63, 66, 72, 61, 62, 63, 62, 64, 60, 114, 62, 62, 77]
+            "samples": [62, 62, 63, 68, 62, 64, 62, 68, 62, 65, 65, 62, 63, 62, 64, 62, 69, 67, 63, 66, 72, 61, 62, 63, 62, 64, 60, 114, 62, 62, 77, 63]
           },
           "last_updated": "2026-09-21"
         },
@@ -2146,7 +2146,7 @@
             "min": 243,
             "max": 349,
             "n": 32,
-            "samples": [252, 247, 250, 264, 260, 251, 349, 250, 268, 248, 257, 266, 250, 254, 248, 262, 246, 273, 264, 254, 263, 284, 258, 248, 254, 249, 260, 243, 309, 250, 249, 324]
+            "samples": [247, 250, 264, 260, 251, 349, 250, 268, 248, 257, 266, 250, 254, 248, 262, 246, 273, 264, 254, 263, 284, 258, 248, 254, 249, 260, 243, 309, 250, 249, 324, 255]
           },
           "last_updated": "2026-09-21"
         },
diff --git a/test/scenarios/light/scenario_peripheral_switch.json b/test/scenarios/light/scenario_peripheral_switch.json
index fc482833..fac351d0 100644
--- a/test/scenarios/light/scenario_peripheral_switch.json
+++ b/test/scenarios/light/scenario_peripheral_switch.json
@@ -660,7 +660,7 @@
             "min": 4,
             "max": 5,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },
@@ -797,7 +797,7 @@
             "min": 4,
             "max": 5,
             "n": 32,
-            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
+            "samples": [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
           },
           "last_updated": "2026-09-21"
         },