Best possible solution:

Land this narrow sentinel-based fix after required CI is green, keeping operational keyring/backend errors fatal while routing corrupt stored credentials to re-auth.

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

Yes. Current main is source-reproducible: corrupt stored token JSON returns a generic decode error and the caller only maps missing keyring entries to re-auth; I did not execute tests because this review is read-only.

Is this the best way to solve the issue?

Yes. The PR uses the narrow sentinel/classifier approach requested by the linked report, so corrupt present tokens can re-auth without masking operational keyring errors.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against fc13b4147e20.

Label changes:

• add P2: A corrupt stored OAuth token can block commands for the affected account, but the fix is localized and the blast radius is limited.
• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool 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: This is a bot-generated ClawSweeper autofix PR, so the external-contributor real behavior proof gate does not apply.

Label justifications:

• P2: A corrupt stored OAuth token can block commands for the affected account, but the fix is localized and the blast radius is limited.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: This is a bot-generated ClawSweeper autofix PR, so the external-contributor real behavior proof gate does not apply.

What I checked:

• Repository policy read: The full target AGENTS.md was read; its PR-review guidance was applied by using gh pr view/gh pr diff without changing branches or editing files. (AGENTS.md:35, fc13b4147e20)
• Current-main bug path: On current main, stored-token JSON decode failures return a generic decode token error rather than a classifiable corrupt-token sentinel. (internal/secrets/token.go:188, fc13b4147e20)
• Current-main auth routing: On current main, the Google API token-source path converts only keyring.ErrKeyNotFound into AuthRequiredError; other GetToken errors remain fatal. (internal/googleapi/client_auth.go:238, fc13b4147e20)
• PR implementation: The PR head adds ErrCorruptStoredToken, wraps both email-token and subject-token JSON decode failures with it, and maps it alongside missing tokens to AuthRequiredError. (internal/secrets/token.go:32, e6f2a1fa7574)
• Caller change: The changed auth path preserves fatal handling for non-sentinel token-store errors while allowing corrupt stored tokens to use the re-auth flow. (internal/googleapi/client_auth.go:239, e6f2a1fa7574)
• Regression coverage: The PR adds tests for corrupt token classification, missing-token separation, subject-token corruption, and caller-level AuthRequiredError mapping. (internal/googleapi/client_more_test.go:169, e6f2a1fa7574)

Likely related people:

• Peter Steinberger: History shows Peter split internal/secrets/token.go, refactored injected Google API auth dependencies, and authored much of the current token-source and keyring code around this path. (role: recent auth/keyring area contributor; confidence: high; commits: 978dd0d14988, 05b75b44cd41, 1531013e50b6; files: internal/secrets/token.go, internal/secrets/store.go, internal/googleapi/client_auth.go)
• Lachlan Donald: Recent history shows Lachlan worked on keyring prompt-storm and token-read behavior adjacent to corrupt or unavailable token storage handling. (role: adjacent auth/keyring contributor; confidence: medium; commits: c58b381c97ad; files: internal/secrets/token.go, internal/secrets/store_more_test.go, internal/cmd/auth_list_helpers.go)

• 🦀 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.

• 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.