feat(menubar): finish opt-in provider colors

[aronchick]

Summary

Add a single Color by provider toggle to the existing menu bar settings. It defaults off and overlays the selected Critters, Meter bars, or Icon & percent style, including each provider's own stacked row. Saved layouts, the existing style picker, and Usage & Spend's Brand/Monochrome artwork remain unchanged.

Provider colors fall back to monochrome during menu tracking, stale data, system Increase Contrast, and the inactive-display contrast preference. Menu close restores color immediately. Accents must pass a 2:1 threshold against conservative reference backgrounds: 25% sRGB gray in dark appearance and 85% in light appearance. This does not sample wallpaper or claim WCAG text conformance. Pace colors keep their separate meaning, and VoiceOver labels do not depend on color.

The maintainer follow-up preserves @aronchick's commits, simplifies the duplicated tint logic, covers standard and vibrant AppKit appearances, adds rendered-pixel and lifecycle regressions, and localizes the one toggle label in every catalog. Thanks @aronchick!

Verification

All Swift runs use source Scripts/test_environment.sh and the native backend.

• swift test --build-system native --jobs 4 -Xswiftc -gnone --no-parallel --filter 'MenuBarProviderColorTests|MenuBarLayoutRendererTests|MergedIconPresentationTests|MenuBarLayoutDisplayOptionsTests|MenuBarPaceColorSettingsTests|SettingsStoreMergeIconStackedTests|StatusMenuAppearanceTests|StatusItemAnimationSignatureTests|PreferencesDocumentTests|LocalizationLanguageCatalogTests|ProviderIconResourcesTests': 192 tests in 12 suites passed.
• Regression runs against the original contribution reproduced contrast/accessibility/lifecycle failures; the final tests pass. A separate red test caught vibrantLight and actual accessibility appearance names before the classifier repair.
• make check: 0 violations, 0 serious, 2,844 Swift files, with repository checks passing. An unchanged process-cleanup fixture timed out on the shared host during an earlier attempt; the retry passed.
• Independent Codex autoreview: no actionable P0–P2 findings.
• swift test --build-system native --jobs 4 -Xswiftc -gnone --no-parallel --filter ProviderArchitectureGatekeeperTests: 48 tests passed. The extracted Warp blink helper retains the same behavior; its exact gatekeeper anchor is updated.
• CODEXBAR_TEST_SUITE_TIMEOUT=600 ./Scripts/test.sh --swift-command <native-wrapper> --direct-workers 4: full inventory of 1,592 selections / 144 groups, with 14,080 discovered methods verified against the direct runtime. The native wrapper forwards test and build to Swift with --build-system native --jobs 4 -Xswiftc -gnone.
• The fail-fast run completed 95 groups. Kimi and Kiro transport retries passed, and a large cost-storage group passed all 12 isolated selections. A Kiro PTY cleanup deadline stopped scheduling later groups; swift test --skip-build --build-system native --jobs 4 -Xswiftc -gnone --no-parallel --filter KiroStatusProbeTests then passed 57 tests.
• The remaining original groups 96–144 were run through Scripts/direct_swift_test_groups.py after revalidating the same complete runtime inventory: 49 groups / 546 selections passed, with 0 failures, 0 retries, 0 timeouts. All original selections are covered on the final head, with 0 unresolved failures. This is aggregate full-suite proof, not a claim that the initial fail-fast command exited successfully.

Synthetic rendered proof

Real renderer and AppKit drawing with fixture values, including pixel assertions in light/dark and fallback cases. This is offscreen proof, not a desktop capture or a live-account/app relaunch.

Refs #3533
Refs #3628
Refs #4294

[clawsweeper]

🦞👀
ClawSweeper picked this up.

Pull request received. I will update this pull request when review starts.

ClawSweeper review complete

ClawSweeper finished reviewing this revision. The review result is being finalized.

View the workflow run.

[clawsweeper]

Codex review: needs real behavior proof before merge. Reviewed October 7, 2026, 10:42 PM ET / October 8, 2026, 02:42 UTC (Revision 4).

ClawSweeper review

What this changes

Adds a default-off Color by provider preference across existing menu-bar styles, with monochrome fallbacks for tracking, stale data, and accessibility states.

Example: Enable Color by provider for Claude

• Before: The menu-bar layout displays “5h 25% W 60%” in monochrome.
• After: The same layout uses Claude’s orange accent in supported light and dark appearances.

Review scores

Measure	Result	What it means
Overall readiness	🦐 gold shrimp (3/6)	Useful implementation with earlier defects resolved and regression completion recorded, but proof still stops at offscreen rendering.
Proof confidence	🦐 gold shrimp (3/6)	Needs stronger real behavior proof before merge: The inspected screenshot exercises the production renderer and button-content application with fixture values in offscreen NSButtons, showing orange output and monochrome fallbacks. It does not exercise the running app’s settings-to-StatusItemController path or native menu-close recovery. Existing-state compatibility is source-verified: absent keys remain false, older imports preserve the preference, and saved layouts are unchanged. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Product

Kind: Feature · Worth it: Yes · Fix scope: Complete
User problem: Users cannot distinguish menu-bar providers by their accent colors.
Reason: One optional preference supplies provider identity cues without changing saved layouts or existing defaults. The later owner-authored follow-up supplies direction for this scope.

Merge readiness

⛔ Blocked before merge - 2 items remain

Keep open; this PR needs real behavior proof before merge. Earlier functional findings are addressed and completed regression coverage is now recorded, but the supplied screenshot still demonstrates only offscreen rendering.

Priority: P3
Reviewed head: 6524e1fdb378f3855dca27a4e972d6b302363da2

Before merge

• Add real behavior proof - Needs stronger real behavior proof before merge: The inspected screenshot exercises the production renderer and button-content application with fixture values in offscreen NSButtons, showing orange output and monochrome fallbacks. It does not exercise the running app’s settings-to-StatusItemController path or native menu-close recovery. Existing-state compatibility is source-verified: absent keys remain false, older imports preserve the preference, and saved layouts are unchanged. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
• Complete next step (P2) - Provide proof from a freshly built native application covering the toggle, provider colors, menu-close restoration, and accessibility fallback. Screenshots, recordings, logs, or live output are acceptable; redact account details, IP addresses, keys, and private endpoints. Updating the PR body triggers a fresh review; otherwise a maintainer can comment @clawsweeper re-review.

Findings

None.

Provenance

• Sources/CodexBar/StatusItemController+MenuTracking.swift: tracking refresh lifecycle: original intent not found (unknown: Feature-history logs and the menu-session refactor were inspected; exact introduction searches failed on missing historical blobs with HTTP 403.)
• Sources/CodexBar/MenuBarLayoutEditor.swift: preview appearance: original intent not found (unknown: The historical content search for the preview appearance failed because the promisor remote returned HTTP 403 for a required blob.)
• Sources/CodexBar/StatusItemController+Animation.swift: Warp blink extraction: original intent not found (unknown: The historical content search failed on a missing blob with HTTP 403; source comparison shows the extracted calculation retains its existing loading condition and fallback.)

Tests

• Low-value test Tests/CodexBarTests/MenuBarProviderColorTests.swift: contrast policy falls back when highlighted stale or under high contrast: Repeats helper guard assertions already covered by rendered fallback, pixel, and lifecycle tests; remove this redundant helper-only test.
• Missing end-to-end proof: The supplied evidence does not show a freshly built application enabling the toggle, coloring native status items, restoring color after menu closure, and responding to Increase Contrast. Reviewer execution of a base-failing/head-passing regression was excluded by the read-only contract.

Agent review details

How this fits together

CodexBar displays provider usage in native macOS menu-bar items. Saved preferences, provider accents, and usage snapshots determine the icons, percentages, and spoken accessibility labels.

flowchart TD
 A[Saved display preferences] --> C[Menu-bar controller]
 B[Provider usage and accent] --> C
 C --> D[Appearance and tracking checks]
 D --> E{Provider color allowed?}
 E -->|Yes| F[Colored native content]
 E -->|No| G[Monochrome native content]

Loading

Technical review

Best possible solution:

Retain the shared optional tint path and establish its settings, native menu tracking, and accessibility behavior in a freshly built application.

Do we have a high-confidence way to reproduce the issue?

Not applicable to a new appearance capability; the inspected screenshot establishes offscreen colored output, while native application behavior remains unproven.

Is this the best way to solve the issue?

Yes at the implementation level: the shared renderer and existing accent palette provide one color source while preserving layouts and defaults.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against 844b0e19bbbb.

Provenance checked

• Sources/CodexBar/MenuBarLayoutRenderer.swift: attachment tinting changes intended behavior with a stated reason (89b42bc: Restore explicit foreground tinting because text attachments draw template images directly; Fix custom layout provider icon rendering #2334 explains the native drawing boundary.)
• Sources/CodexBar/MenuBarLayoutRenderer.swift: pace-token colors keeps the original intent (Add optional colors for menu bar quota pace #3429: Optional green/red pace colors communicate reserve versus usage ahead of pace while neutral values and existing layouts retain their behavior.)
• Sources/CodexBar/StatusItemController+MenuBarLayout.swift: stacked provider rows keeps the original intent (feat(menu): add stacked two-provider icon style #3628: Each provider uses its first effective layout row while retaining the existing status item, placement, independent updates, and saved preferences.)

Testing

Proof path: in-process harness. Added test files: 45.

Security

None.

Evidence

What I checked:

• Applicable repository policy: Read the complete root AGENTS.md; no nested AGENTS.md or maintainer-notes directory was found. Applied native-bundle validation, isolated settings, and provider-separation guidance. Builds and tests were not executed under the read-only contract. (AGENTS.md:1, 6524e1fdb378)
• Pinned scope and current-main overlap: Read the pinned feature hunks and independently compared tree objects: 155 of 199 introduced paths have identical blobs on fetched main. Current main lacks the provider-color preference; the existing pace-color setting and Usage & Spend artwork address different behavior. Endpoint differences in unrelated current-main work were not attributed to this PR. (Sources/CodexBar/PreferencesMenuBarPane.swift:58, 6524e1fdb378)
• Later owner direction: The owner declined a separate color mode on CodexBar menu bar color mode #3533, then authored the later follow-up explicitly adopting one default-off preference with contrast and tracking fallbacks. That newer recorded direction covers this implementation. (fc046bfd8aae)
• Prior functional findings resolved: The tint policy rejects highlighted, stale, high-contrast, and accessibility states and checks the documented 25%/85% gray backgrounds. Tracking start/end refresh content; classification follows the tracking session and merged refresh bypasses tracking deferral. No earlier finding remains actionable. (Sources/CodexBar/MenuBarLayoutRenderer.swift:973, 6524e1fdb378)
• Prepared screenshot inspected: Inspected the downloaded image at https://github.com/user-attachments/assets/8334eaec-6257-4f96-928b-d6cc17d8ed47. It visibly shows orange Claude output and monochrome fallback panels and explicitly identifies itself as synthetic offscreen proof. The source constructs NSButtons with fixture values rather than exercising the shipped settings-to-status-item path. (Tests/CodexBarTests/MenuBarProviderColorTests.swift:348, 6524e1fdb378)
• Existing-state compatibility: An absent defaults key resolves to false. Import only changes the preference when the optional key exists, preserving older documents and current values. The change neither rewrites saved layouts nor requires a stored-data migration. (Sources/CodexBar/SettingsStore+PreferencesDocument.swift:39, 6524e1fdb378)

Likely related people:

• steipete: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)
• ysyyork: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)
• Jon Brown: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)

Review metrics

Metric	Value	Why it matters
Current-main overlap	155 of 199 introduced paths have identical current-main blobs	Much of the headline branch size consists of already-integrated upstream work.
Provider-color code growth	Production +208/-27; tests +549/-3 versus fetched main	The shared tint path and rendering regressions justify the bounded growth; production includes 46 localization lines.

Labels

Label changes:

No label changes.

Label justifications:

• P3: This is an optional appearance improvement with no reported broken core workflow.
• rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦐 gold shrimp and patch quality is 🐚 platinum hermit.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs stronger real behavior proof before merge: The inspected screenshot exercises the production renderer and button-content application with fixture values in offscreen NSButtons, showing orange output and monochrome fallbacks. It does not exercise the running app’s settings-to-StatusItemController path or native menu-close recovery. Existing-state compatibility is source-verified: absent keys remain false, older imports preserve the preference, and saved layouts are unchanged. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
• proof: 📸 screenshot: Contributor real behavior proof includes screenshot evidence.

Rank-up moves

Optional improvements that raise the rating; they are not merge blockers.

• Add redacted native application screenshots, a recording, or runtime output showing the toggle, provider colors, menu-close restoration, and accessibility fallback.
• Remove the redundant helper-only fallback test.

Rating scale

6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior.

Workflow

ClawSweeper edits this one comment on every review. Comment @clawsweeper re-review for a fresh review only; repair and merge need explicit maintainer commands such as @clawsweeper autofix or @clawsweeper automerge.

History

Review history (3 earlier review cycles)

• reviewed 2026-10-07T07:01:10.520Z sha bb64180 :: needs real behavior proof before merge. :: [P2] Include system Increase Contrast in the tint fallback | [P2] Refresh provider colors after clearing open-menu tracking | [P2] Evaluate contrast against the supported menu-bar background
• reviewed 2026-10-08T01:04:01.696Z sha fc046bf :: needs real behavior proof before merge. :: none
• reviewed 2026-10-08T01:49:23.102Z sha 6524e1f :: needs real behavior proof before merge. :: none

[steipete]

Thanks @aronchick! Merged in fe95bfd with your commits preserved. Color by provider ships as a single opt-in toggle on the menu bar layout (default off), tinting each provider slot with its brand colour on top of the active layout and icon style; the maintainer follow-ups made VoiceOver labels independent of colour, repaired the menu-close fallback, and fixed the contrast failure with rendered-colour regressions for both appearances. Ships in the next release.