D-RPF-0: CE64 field predicate over resident NodeRow bytes (test-only probe)

.claude/board/STATUS_BOARD.md

@@ -4,7 +4,7 @@ Plan: `.claude/plans/2026-10-08-resident-projection-fold-mask-v1.md`. Mask = adm
 
 | D-id | scope | status | gate / falsifier |
 |---|---|---|---|
-| **D-RPF-0** | Pearl3 / Epi5 / mantissa predicate over `MaterializedEdges` via `MatchFacet16Strided`, threshold as ≤ n+1 patterns | Queued | equals CE64 accessors; wrong-bit care disagrees; unnamed fields silent; both windows stay inside the 32 B tenant |
+| **D-RPF-0** | Pearl3 / Epi5 / mantissa predicate over `MaterializedEdges` via `MatchFacet16Strided`, threshold as ≤ n+1 patterns | In PR (test-pinned, `mask-risc/tests/resident_ce64_predicate.rs`) | equals CE64 accessors; wrong-bit care disagrees; unnamed fields silent; both windows stay inside the 32 B tenant |
 | **D-RPF-1** | Law tables → per-recipe ternary pattern set (≤ 24 × 8), run under an admission plane (class, rail, generation, provenance) | Queued | equals `measure_declared`; dropping one code loses exactly its rows; removing the admission plane admits a v1 row |
 | **D-RPF-2** | Census of eligibility and read-set classes on a real population | Queued | classes/rows reported; near 1 drops D-RPF-3; no population-sized seen-set |
 | **D-RPF-3** | Exact read-set dedup per CE64 instruction, key incl. handle context | Queued (after D-RPF-2) | removing cohort from key changes a result; dedup-then-fold equals per-row for Count/Sum/Avg |

.claude/board/entries/2026-10-08-d-rpf-0-resident-ce64-predicate.md

@@ -0,0 +1,40 @@
+# 2026-10-08 — D-RPF-0: CE64 field predicate over resident `NodeRow` bytes
+
+**Status:** MEASURED, TEST-PINNED
+(`crates/lance-graph-mask-risc/tests/resident_ce64_predicate.rs`, 6 tests;
+`cargo test -p lance-graph-mask-risc --test resident_ce64_predicate`).
+Test-only: the library is unchanged; `causal-edge` and `lance-graph-contract`
+are dev-dependencies of mask-risc.
+
+## Claim and result
+
+A Pearl3, Inference (raw 4-bit nibble) or Epi5 predicate on any of the four
+`CausalEdge64` words in `ValueTenant::MaterializedEdges` runs in place through
+`Pred::MatchFacet16Strided` over two 16-byte windows (edges 0,1 and 2,3), with
+no extracted lane. Executor and oracle agree, and both equal the CE64 named
+accessors, row by row, for every value of every field on every edge
+(3 tiles + 37 rows). `raw5 >= t` is a union of at most 6 patterns and equals
+the accessor for all 32 thresholds on all four edges.
+
+The predicate is built from `isa::Field::span()`; the oracle reads the CE64
+accessors, which never consult that table. Offsets come from
+`ValueTenant::value_offset()`.
+
+## Falsifiers
+
+- can fire: a care shifted up by one bit disagrees with the accessor for some
+  value of every field;
+- can stay silent: rewriting every byte of every row except the cared field
+  leaves the mask unchanged;
+- anti-vacuity: each field has a value admitting a strict, non-empty minority
+  (< 1/3) of rows. It fired on the first run: the fixture generator (`>> 11`
+  LCG) left bits 53..63 zero in every word, so Epi5 was constant. Replaced by
+  SplitMix64.
+
+Disable runs (anchor asserted, each red): window `k` instead of `k / 2` (4),
+half always 0 (3), threshold without the equality pattern (1), silent-arm noise
+keeping none of the base field (1).
+
+## OPEN
+
+None for D-RPF-0. D-RPF-1 adds the admission plane in front of these patterns.

.claude/board/entries/README.md

@@ -25,7 +25,7 @@ index row, (3) no duplicate entry id. Checks 1 and 2 are deliberately
 opposite directions; the stranding this convention prevents shows up in
 exactly one of them, never both.
 
-271 entries, 2026-08-06 .. 2026-10-08.
+272 entries, 2026-08-06 .. 2026-10-08.
 
 | date | entry id | finding | file |
 |---|---|---|---|
@@ -36,6 +36,7 @@ exactly one of them, never both.
 | 2026-10-08 | `moore-nars16-isa-visible-representation` |  | [2026-10-08-moore-nars16-isa-visible-representation.md](2026-10-08-moore-nars16-isa-visible-representation.md) |
 | 2026-10-08 | `D-MOORE-NARS-0` |  | [2026-10-08-moore-nars-0-recipe-learning-gomoku.md](2026-10-08-moore-nars-0-recipe-learning-gomoku.md) |
 | 2026-10-08 | `hhtl-nars-moore-value-tenants` |  | [2026-10-08-hhtl-nars-moore-value-tenants.md](2026-10-08-hhtl-nars-moore-value-tenants.md) |
+| 2026-10-08 | `D-RPF-0` |  | [2026-10-08-d-rpf-0-resident-ce64-predicate.md](2026-10-08-d-rpf-0-resident-ce64-predicate.md) |
 | 2026-10-08 | `coresearch-ce64-moore-masking-wiring` |  | [2026-10-08-coresearch-ce64-moore-masking-wiring.md](2026-10-08-coresearch-ce64-moore-masking-wiring.md) |
 | 2026-10-08 | `ce64-isa-register-contract` |  | [2026-10-08-ce64-isa-register-contract.md](2026-10-08-ce64-isa-register-contract.md) |
 | 2026-10-07 | `tinker-janus-fold-harvest` | TinkerPop bulk = K (GroupReduce), ONE_BULK = S, GValue pinning = bundle invalidation; JanusGraph slice = OrderedLaneWitness→Range (30–38×); no new V4 op | [2026-10-07-tinker-janus-fold-harvest.md](2026-10-07-tinker-janus-fold-harvest.md) |

.claude/plans/2026-10-08-resident-projection-fold-mask-v1.md

@@ -51,7 +51,7 @@ between them. Where a hop has no primitive, the phase stops and files the gap.
 
 ## Checklist
 
-- [ ] **D-RPF-0** — Resident CE64 field predicate in place
+- [x] **D-RPF-0** — Resident CE64 field predicate in place (`lance-graph-mask-risc/tests/resident_ce64_predicate.rs`; entry `2026-10-08-d-rpf-0-resident-ce64-predicate.md`)
 - [ ] **D-RPF-1** — Law-table → pattern-set compiler (recipe eligibility as a mask)
 - [ ] **D-RPF-2** — Class census: how many distinct eligibility and read-set classes a real population has
 - [ ] **D-RPF-3** — Read-set dedup for CE64 instructions (hotspot dedup, exact)

.github/workflows/rust-test.yml

@@ -436,8 +436,16 @@ jobs:
         run: cargo test --manifest-path crates/sigma-tier-router/Cargo.toml
       - name: Run neural-debug tests (previously ungated)     # 11 green
         run: cargo test --manifest-path crates/neural-debug/Cargo.toml
+      # opt-level 3 for THIS package only, debug stays 0 (manifest). The
+      # crossword_real_words_probe example tests (D-PUZZLE-0, 2026-10-07) run a
+      # solver search that takes 810 s at opt-level 0 and pushed this job past
+      # its 30-minute limit on main (run 37766947128, cut off in `hydrate`).
+      # Measured locally at opt-level 3: 183 s, 4m08 including the compile.
+      # `--config` keeps every other crate on the shared opt-level-0 cache.
       - name: Run shader-driver tests (previously ungated)    # 107 + 2 green
-        run: cargo test --manifest-path crates/cognitive-shader-driver/Cargo.toml
+        run: >-
+          cargo test --manifest-path crates/cognitive-shader-driver/Cargo.toml
+          --config 'profile.dev.package.cognitive-shader-driver.opt-level=3'
       # The default run above skips these targets: `w2_differential` and
       # `mailbox_cutover` are `cfg(feature = "mailbox-thoughtspace")`, and the
       # StreamDto circuit probe needs `with-engine` for `StreamDto`.

crates/lance-graph-mask-risc/Cargo.toml

@@ -12,3 +12,8 @@ description = "The one mask-RISC evaluator/fuser above ndarray::simd — a plan
 ndarray = { path = "../../../ndarray", default-features = false, features = ["std"] }
 
 [dev-dependencies]
+# D-RPF-0 (`tests/resident_ce64_predicate.rs`): the CE64 accessors are the
+# oracle and the contract supplies the tenant offsets. Test-only — the library
+# still names neither.
+causal-edge = { path = "../causal-edge" }
+lance-graph-contract = { path = "../lance-graph-contract" }