Fix slow Claude CLI failure refreshes

[steipete]

Summary

• stop the Claude CLI probe before /status when /usage already returned a terminal error or loading stall
• preserve /status identity enrichment for successful usage snapshots
• stabilize the new cost-window test around local calendar days
• document the Claude refresh improvement in the changelog

Proof

• swift test --filter TTYIntegrationTests
• swift test --filter MenuCardCostComparisonTests
• swift test --filter StatusProbeTests
• swift test --filter ClaudeCLITimeoutRetryTests
• make check
• make test — 47/47 shards passed
• autoreview — no actionable findings

Live Claude CLI failure timings improved from 19.22s / 7.68s / 19.43s to 7.77s / 7.33s / 8.04s, removing the redundant interactive identity round trip. The freshly packaged app also completed authenticated Claude Web fetches successfully and rendered session, weekly, daily, and extra-usage windows in the menu.

[clawsweeper]

Codex review: needs maintainer review before merge. Reviewed July 4, 2026, 2:20 PM ET / 18:20 UTC.

Summary
This PR adds a pre-/status validation step for Claude CLI /usage terminal errors/loading stalls, updates focused TTY and cost-window tests, and adds a changelog note.

Reproducibility: yes. from source inspection, though I did not execute tests in this read-only review. Current main captures /status before parsing /usage, and the PR's TTY fixture demonstrates the subscription-notice path that should stop before /status.

Review metrics: none identified.

Merge readiness
Overall: 🦞 diamond lobster
Proof: 🦞 diamond lobster
Patch quality: 🦞 diamond lobster
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Next step before merge

• No ClawSweeper repair lane is needed; this owner-authored PR has no actionable patch findings and should proceed through normal maintainer merge/check handling.

Security
Cleared: The diff does not change dependencies, secrets handling, CI, release scripts, network permissions, or supply-chain execution paths.

Review details

Best possible solution:

Land the small provider fix after normal required checks, keeping the pre-status validation and focused TTY regression coverage.

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

Yes from source inspection, though I did not execute tests in this read-only review. Current main captures /status before parsing /usage, and the PR's TTY fixture demonstrates the subscription-notice path that should stop before /status.

Is this the best way to solve the issue?

Yes. Moving the existing usage-error/loading-stall checks before /status is the narrow maintainable fix and preserves identity enrichment for successful snapshots.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 8b417a066cb1.

Label changes

Label changes:

• add P2: The PR fixes a real Claude-provider refresh latency issue with limited blast radius and focused tests.
• add proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes focused test commands plus live Claude CLI timing output and freshly packaged app validation for the changed behavior.
• add rating: 🦞 diamond lobster: Overall readiness is 🦞 diamond lobster; proof is 🦞 diamond lobster and patch quality is 🦞 diamond lobster.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes focused test commands plus live Claude CLI timing output and freshly packaged app validation for the changed behavior.

Label justifications:

• P2: The PR fixes a real Claude-provider refresh latency issue with limited blast radius and focused tests.
• rating: 🦞 diamond lobster: Overall readiness is 🦞 diamond lobster; proof is 🦞 diamond lobster and patch quality is 🦞 diamond lobster.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (live_output): The PR body includes focused test commands plus live Claude CLI timing output and freshly packaged app validation for the changed behavior.
• proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes focused test commands plus live Claude CLI timing output and freshly packaged app validation for the changed behavior.

Evidence reviewed

What I checked:

• Repository policy read: AGENTS.md was read fully; its guidance favors focused provider/parser tests and avoiding live provider probes that may trigger Keychain prompts unless explicitly requested. (AGENTS.md:23, 8b417a066cb1)
• Current main behavior: Current main captures /status before calling parse(text:statusText:), so terminal /usage errors and loading stalls still pay the identity round trip. (Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift:110, 8b417a066cb1)
• Existing parser boundary: Current main already treats usage errors and loading stalls as parse-time terminal conditions, which makes moving those checks before /status a narrow refactor of existing behavior. (Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift:432, 8b417a066cb1)
• PR implementation: The PR head validates the /usage text before capturing /status, while leaving /status identity enrichment in place for valid usage snapshots. (Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift:112, 1018f0eff544)
• Focused regression coverage: The TTY regression test logs commands sent to the fake Claude CLI and asserts /usage is sent while /status is not sent for the subscription-notice failure path. (Tests/CodexBarTests/TTYIntegrationTests.swift:96, 1018f0eff544)
• PR proof: The PR body reports focused Swift test filters, make check, make test with 47/47 shards passed, live Claude CLI failure timing improvement, and a freshly packaged app run showing Claude Web menu data. (1018f0eff544)

Likely related people:

• steipete: Current blame and recent history show Peter Steinberger owning the current ClaudeStatusProbe fetch/error path and this focused PR's commits. (role: recent area contributor; confidence: high; commits: f84a171bca1b, 27f99eb7f22c, 258f08e2e1b3; files: Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift, Sources/CodexBarCore/Providers/Claude/ClaudeUsageFetcher.swift)
• Yuxin-Qiao: Commit history ties this person to Claude CLI subscription usage output handling, which is one terminal failure path affected by this PR. (role: adjacent feature contributor; confidence: medium; commits: 126dc9238715; files: Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift, Tests/CodexBarTests/StatusProbeTests.swift)
• rohitjavvadi: Commit history ties this person to Claude usage loading stall handling, the other terminal condition moved before /status by this PR. (role: adjacent feature contributor; confidence: medium; commits: 354e0b6dd19e; files: Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift, Tests/CodexBarTests/StatusProbeTests.swift)
• ratulsarna: Commit history ties this person to earlier Claude PTY/session handling that this probe path still uses. (role: adjacent owner; confidence: medium; commits: a86e240ee9e7; files: Sources/CodexBarCore/Providers/Claude/ClaudeStatusProbe.swift, Sources/CodexBarCore/Providers/Claude/ClaudeCLISession.swift)

What the crustacean ranks mean

• 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
• 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
• 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
• 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
• 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
• 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
• 🌊 off-meta tidepool: rating does not apply to this item.

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.

How this review workflow works

• 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.