train: land 4081 4082 4087 4089 4084 4086 4092 4091 4093

[steipete]

Merge train: one CI run for 9 green lane PRs, each kept as its own commit (rebase merge).

• fix(config): treat blank config files as absent #4081
• fix(menu-bar): validate preserved status item positions #4082
• fix(codex): refresh validated snapshots during catch-up #4087
• fix(keychain): bound validation stalls and recover Claude caches #4089
• fix(antigravity): preserve grouped OAuth quota summaries #4084
• fix(kimi): clarify stale CLI credential recovery #4086
• fix(cost): compact and reuse Claude cache artifacts #4092
• fix(providers): clarify Kimi and z.ai quota availability #4091
• fix(grok): preserve local token totals in Usage & Spend #4093

[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: blocked before merge. Reviewed September 28, 2026, 3:40 AM ET / 07:40 UTC.

ClawSweeper review

What this changes

Combines nine fixes for configuration, menu placement, provider quotas, credential recovery, and usage history into separate commits with tests and documentation.

Merge readiness

⛔ Blocked before merge - 1 item remains

Keep open. Current main lacks the nine fixes carried by this owner-authored merge train. The reviewed source and regression coverage establish no concrete introduced blocker.

Priority: P1
Reviewed head: 58f3f0a899ade89c59f3a26f1f143e265b6c6a9c

Review scores

Measure	Result	What it means
Overall readiness	🐚 platinum hermit (4/6)	Focused regression coverage, an upgrade fixture, and linked before/after evidence support a broad but reviewable train.
Proof confidence	🐚 platinum hermit (4/6)	Not applicable: The owner-authored train is exempt from the external contributor proof gate. Linked PR evidence covers isolated CLI config behavior, native Kimi rendering, and a version-3 cache rebuild; Antigravity OAuth was exercised with a synthetic transport, and no separate integrated live run is claimed. The cache upgrade fixture supports existing-state compatibility.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Verification

Check	Result	Evidence
Real behavior	Not applicable	Not applicable: The owner-authored train is exempt from the external contributor proof gate. Linked PR evidence covers isolated CLI config behavior, native Kimi rendering, and a version-3 cache rebuild; Antigravity OAuth was exercised with a synthetic transport, and no separate integrated live run is claimed. The cache upgrade fixture supports existing-state compatibility.
Evidence reviewed	7 items	Introduced train: The pinned main-to-head diff contains nine commits and changes 67 files; the config loader, Grok publication, and Claude cache changes are absent from the pinned main. Credential boundary: Timed-out signature validation returns an inconclusive result; the Claude recovery path checks profile-scoped, unexpired memory before persisting it with its recorded owner. OAuth fallback: Antigravity tries measured grouped quotas with a two-second timeout and retains model-quota fallback, selected-account identity, cancellation, and authentication failure handling.
Findings	None	None.
Security	None	None.

How this fits together

CodexBar reads configuration, credentials, provider responses, and local session files to build usage snapshots. Those snapshots feed the menu bar, widget, CLI, and Usage & Spend views.

flowchart LR
 A[Config and credentials] --> B[Provider probes]
 C[Local session files] --> D[Cost scanner]
 B --> E[Usage snapshots]
 D --> E
 E --> F[Menu, widget, CLI, and dashboard]

Loading

Before merge

• Resolve review confidence - ClawSweeper must reach high confidence before merge readiness is known.

Agent review details

Security

None.

Review metrics

Metric	Value	Why it matters
Train scope	9 commits, 67 files	The branch combines separately attributable fixes across several provider and app paths.
Production and test delta	production +368/−394; tests +1602/−196	The fixes reduce net production lines while adding focused regression coverage.

Technical review

Best possible solution:

Land the separately attributable fixes through the train after its ordinary exact-head checks, preserving the tested cache upgrade and credential guards.

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

Not applicable as one issue: this train combines nine fixes with focused reproduction or regression evidence in their linked PRs. This read-only review did not execute those paths.

Is this the best way to solve the issue?

Yes. The train retains separate commits and uses existing snapshot, parser, and cache boundaries; no narrower repair is indicated by the reviewed diff.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against a5252e24c844.

Labels

Label changes:

• add P1: One included fix addresses native validation stalls that can block background provider refresh.
• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: The owner-authored train is exempt from the external contributor proof gate. Linked PR evidence covers isolated CLI config behavior, native Kimi rendering, and a version-3 cache rebuild; Antigravity OAuth was exercised with a synthetic transport, and no separate integrated live run is claimed. The cache upgrade fixture supports existing-state compatibility.

Label justifications:

• P1: One included fix addresses native validation stalls that can block background provider refresh.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: The owner-authored train is exempt from the external contributor proof gate. Linked PR evidence covers isolated CLI config behavior, native Kimi rendering, and a version-3 cache rebuild; Antigravity OAuth was exercised with a synthetic transport, and no separate integrated live run is claimed. The cache upgrade fixture supports existing-state compatibility.

Evidence

What I checked:

• Introduced train: The pinned main-to-head diff contains nine commits and changes 67 files; the config loader, Grok publication, and Claude cache changes are absent from the pinned main. (Sources/CodexBarCore/Config/CodexBarConfigStore.swift:35, 58f3f0a899ad)
• Credential boundary: Timed-out signature validation returns an inconclusive result; the Claude recovery path checks profile-scoped, unexpired memory before persisting it with its recorded owner. (Sources/CodexBarCore/KeychainAccessPreflight+ValidationMemo.swift:44, 58f3f0a899ad)
• OAuth fallback: Antigravity tries measured grouped quotas with a two-second timeout and retains model-quota fallback, selected-account identity, cancellation, and authentication failure handling. (Sources/CodexBarCore/Providers/Antigravity/AntigravityRemoteUsageFetcher.swift:140, 58f3f0a899ad)
• Cache upgrade: The artifact schema advances to version 4; a focused test seeds version 3 rows, rebuilds from transcripts, and checks that totals and derived history remain equal. (Tests/CodexBarTests/CostUsageClaudeWriteAmplificationTests.swift:190, 58f3f0a899ad)
• Prior cache work: Main-branch history connects Daniel JB Clark to decoded Claude artifact reuse before this train; this is routing context, not an introduction claim for a particular line. (Sources/CodexBarCore/Vendored/CostUsage/CostUsageClaudeCache.swift, 741f5cbe1792)
• Related proof: The linked source PRs remain open. Their bodies report an isolated blank-config CLI before/after run, synthetic OAuth and credential regressions, a version-3 cache upgrade test, and a native Kimi rendering; the train body does not claim a separate integrated live run.

Likely related people:

• Peter Steinberger: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)
• Daniel JB Clark: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)
• Lam,Yiu Fung: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)

Rating scale

Score	Internal tier	Crab rank	Meaning
6/6	S	🦀 challenger crab	Exceptional readiness
5/6	A	🦞 diamond lobster	Very strong readiness
4/6	B	🐚 platinum hermit	Good normal PR; ordinary maintainer review
3/6	C	🦐 gold shrimp	Useful, but confidence is limited
2/6	D	🦪 silver shellfish	Proof or implementation needs work
1/6	F	🧂 unranked krab	Not merge-ready
N/A	NA	🌊 off-meta tidepool	Rating does not apply

Overall follows the weaker of proof and patch quality.
Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

Workflow

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.