docs: document completion and custom endpoint settings keys (+ missing_docs bookkeeping for v0.2026.08.26) - #657
Conversation
Add warp_completions_enabled, native_shell_completions_enabled, and agents.custom_endpoints to the all-settings reference for the v0.2026.08.26.17.59.stable_01 missing_docs drift-watch triage. Fold in surface-map internal factory routes, changelog verdicts, release marker, and snapshot regen. Co-Authored-By: Oz <oz-agent@warp.dev>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
@warp-factories[bot] I'm starting a first review of this pull request. You can view the conversation on Warp. I completed the review and no human review was requested for this pull request. Comment Powered by Oz |
|
This PR was generated with Warp. Comment   |
![warp-for-oss[bot]](https://avatars.githubusercontent.com/in/3450139?s=60&v=4){kind=small}
There was a problem hiding this comment.
Overview
This PR updates the all-settings reference with completion and custom inference endpoint settings, and records missing-docs drift-watch bookkeeping for the release. The custom endpoint row and internal bookkeeping look consistent, but the completion settings documentation appears to publish dogfood-gated behavior as generally available.
Concerns
- The completion settings rows should either be deferred until the native completions feature is GA or explicitly state the availability gate. The diff records
NativeShellCompletionsas dogfood, and the runtime forces Warp-only completions when that feature flag is off, ignoring both completion toggles for most ungated users.
Verdict
Found: 0 critical, 1 important, 0 suggestions
Request changes
Comment /warp-agent-review on this pull request to retrigger a review (up to 3 times on the same pull request).
Powered by Oz
| * `warp_completions_enabled` — Whether Warp's built-in completions are shown for shell commands. When enabled, **Settings** > **Features** > **Terminal Input** also exposes "Open completions menu as you type" as a nested control (the same knob as `completions_open_while_typing`). Type: boolean. Default: `true`. | ||
| * `native_shell_completions_enabled` — Whether your shell's own completions are used for shell commands. Native shell completions are not generated as you type; press `Tab` (or your completions keybinding) to fetch them. Type: boolean. Default: `true`. |
There was a problem hiding this comment.
NativeShellCompletions as dogfood-gated and the runtime forces WarpOnly when that flag is off. Public settings docs should either defer these rows until the feature is GA or state the availability gate so stable users do not try settings that are ignored.
…ges map target Second, exhaustive pass over the same release. Adds ledger rows for the surface-scoped slash command finding (shipped as its own feature PR), the 81 --weak-coverage findings, and an itemized re-verification of the 23 potentially_stale_docs findings that supersedes the earlier bulk row. Retargets the KittyImages map entry: it pointed at full-screen-apps.mdx, which documents the Kitty keyboard protocol, but the Kitty image protocol is documented on files-and-links.mdx. Co-Authored-By: Warp <agent@warp.dev>
hongyi-chen
changed the title
…apshot Drift-watch re-run after v0.2026.08.26.17.59.stable_01 bookkeeping: defer the new dogfood flag and regenerate the surface snapshot so the next --diff run is clean of medium findings.
…utes Address docs#657 review: state that warp/native completion toggles only take effect under the Native shell completions Preview. Map the four factory /costs/* API routes as internal, prune dead /metrics/* map entries, record the dogfood→preview NativeShellCompletions deferral, and refresh the surface snapshot.
Extend docs#657 bookkeeping for the v0.2026.08.26 release re-run:
map GET/POST /factory-inbox* and POST /factory/{uid}/learning as
internal (Gate 0 — unreleased Factory REST), record ledger verdicts,
refresh last_release marker timestamp and surface snapshot.
Record Gate 0 deferral for the new preview history-search ranking flag, add it to the surface-map ignore list, and refresh the surface snapshot.
…handoff route
Exhaustive drift-watch re-run for v0.2026.08.26.17.59.stable_01.
- Ignore the new ShellWidgetHandoff preview flag (Gate 0: not GA).
- Map POST /factory/{uid}/integrations/linear/handoff -> internal (Gate 0:
unreleased /factory namespace).
- Record this run's verdicts, including the corrected staleness verdict for
the three live "Warp AI credits" usages.
- Refresh surface_snapshot.json.
Co-Authored-By: Warp <agent@warp.dev>
2 participants
What this feature does
Three always-on settings were missing from the all-settings reference (
terminal.input.warp_completions_enabled,terminal.input.native_shell_completions_enabled,agents.custom_endpoints). This PR documents those keys and carries the missing_docs drift-watch bookkeeping forv0.2026.08.26.17.59.stable_01.The completion toggles are additionally gated by the Native shell completions Preview flag: when that flag is off, runtime forces Warp-only completions and the Settings widgets are hidden. The all-settings rows now state that availability gate so stable users do not treat the keys as generally effective.
Summary
terminal.input.warp_completions_enabledandterminal.input.native_shell_completions_enabledunder[terminal.input], including the Preview availability gate.agents.custom_endpointsunder[agents].v0.2026.08.26.17.59.stable_01: surface map, changelog ledger, release marker, surface snapshot./costs/*routes-> internal(replacing removed/metrics/cost-by-pr-sizeand/metrics/top-prsentries).GET /factory-inbox,POST /factory-inbox/notifications/{uid}/dismiss, andPOST /factory/{uid}/learning-> internal(2026-09-01 re-run).ShellWidgetHandoff(new preview flag), mapPOST /factory/{uid}/integrations/linear/handoff -> internal, record the corrected "Warp AI credits" staleness verdict (fixed in docs(inference): call the inference bucket "AI credits", not "Warp AI credits" #674), and refresh the snapshot.Changes
src/content/docs/terminal/settings/all-settings.mdx— three settings rows; completion rows note Preview gate..agents/skills/missing_docs/references/feature_surface_map.md— settings mappings; factory costs, inbox, learning, and Linear-handoff routes internal; dead metrics paths pruned; KittyImages retarget;ShellWidgetHandoffignore entry..agents/skills/missing_docs/references/changelog_decisions.md— release triage verdicts (rejections included)..agents/skills/missing_docs/references/last_release_processed.jsonandsurface_snapshot.json— release marker + snapshot regen.Content design plan
Short form (update existing all-settings reference):
settings.toml(or reading the all-settings reference) who wants to turn Warp vs native shell completions on or off, or inspect the custom inference endpoints collection key.2026-09-02 exhaustive re-run
Re-audited every category at every severity against
mainmerged into this branch. Both modes exited 0 with no skipped audits andunaccounted: none. After this PR's changes, every coverage bucket is clean:Extra passes beyond the audit defaults:
--weak-coverage(82 findings) — the same 81 mapped-page keyword misses verified on 2026-08-26 and 2026-08-28, plusShellWidgetHandoff. No new gap./usageship undocumented). All 44 GUI-visible commands instatic_commands/commands.rsappear onagents/capabilities/slash-commands.mdxexcept/docker-sandbox, which is deliberately mapped-> internal.NativeShellCompletions,ShellWidgetHandoff, andHistorySearchRankingV2are all still inPREVIEW_FLAGS;AIMemoriesandProviderCommandare still non-GA (theirgated:entries stay deferred); the released OpenAPI still has zero/factorypaths.Warp AI creditsusages as telemetry names or historical references. Corrected and fixed in docs(inference): call the inference bucket "AI credits", not "Warp AI credits" #674.Deferred findings
/factory/*and/factory-inbox/*(including/costs/*, issue-tracker issues, Slack connection-test greeting, Linear handoff, learning, inbox list/dismiss) — Gate 0: namespace excluded from the released OpenAPI until the Factory REST API is public. Mapped-> internal.ShellWidgetHandoff(preview, 2026-09-02) — Gate 0: not GA. Hands Ctrl+R history search to the shell's own fzf/atuin widget. Ignore-list entry; re-check on GA.HistorySearchRankingV2(preview) — Gate 0: not GA. Ignore-list entry; re-check on GA.NativeShellCompletions(preview) — Gate 0: not GA. Settings knobs documented with the Preview gate; flag stays on the ignore list.LrcActivitySignal(dogfood) — Gate 0: tracked by snapshot only.agents.warp_agent.other.usage_display_unit(dogfood / PricingTransparency) — Gate 0: re-check on promotion.gated:AIMemories) andoz provider(gated:ProviderCommand) — Gate 0: research preview / dogfood. Auto-surface when their flags go GA.#15515,#15258,#15424,#15395,#15365,#15340,#15148,#15346,#13967,#15221,#15505, Factories EA bullets) — recorded as no docs needed in the ledger. The singleoz_updatesbullet for this release (#15505, handoff snapshot fix) is a bug fix restoring intended behavior.style_lint; includes the knownai creditsbucket-name false positives and the openai creditspolicy question in the ledger's Notes section, which needs a human docs-owner decision.Validation
npm ci && npm run buildpasses. Re-running the audit after these changes leaves only the already-decided changelog rows and the style_lint-owned terminology hits.Reviewers
Resolved owners for the settings and flag sources (
app/src/settings/input.rs,app/src/settings/ai.rs,crates/warp_features/src/lib.rs) →warpdotdev/oss-maintainers. Factory handler paths are unresolved in warp-server's ownership files (no entry coversrouter/handlers/public_api/), so the fallbackdannyneiraapplies./cc @dannyneira