[codex] Fix Kimi subscription quota rows

[skyzer]

Summary

• Update the Kimi membership stats RPC from GetSubscriptionStat to the current GetSubscriptionStats endpoint so the Monthly subscription quota can populate again.
• Parse ratelimitCode7d from Kimi membership stats and expose it as a separate Code 7-day extra quota row.
• Carry all Kimi extra quota rows into widget snapshots and allow Kimi widgets to render four persisted rows.

Root Cause

Kimi now returns 404 for MembershipService/GetSubscriptionStat. The live membership quota page calls MembershipService/GetSubscriptionStats, which returns subscriptionBalance plus 5-hour / 7-day Code quota ratios. CodexBar was silently dropping the monthly enrichment and had no model field for the subscription 7-day Code quota.

Validation

• git diff --check
• swiftc -parse on the touched source files
• swift test --filter KimiProviderTests could not run locally because this machine's Command Line Tools install fails before compilation with xcrun: error: unable to lookup item 'PlatformPath' from command line tools installation.

[clawsweeper]

Codex review: needs real behavior proof before merge. Reviewed July 6, 2026, 9:02 AM ET / 13:02 UTC.

Summary
The PR updates Kimi membership quota parsing to use GetSubscriptionStats, adds a ratelimitCode7d quota row, persists all Kimi extra quota rows to widgets, raises Kimi widget row limits to four, and extends parser/widget tests.

Reproducibility: no. high-confidence live reproduction is attached. Source inspection confirms current main still calls the endpoint the PR reports as returning 404 and only persists one Kimi extra widget row, but I did not run a real Kimi account probe.

Review metrics: 2 noteworthy metrics.

• Changed surface: 9 files, +101/-23. The patch spans provider parsing, widget snapshot propagation, widget row limits, tests, and changelog text, so provider and widget behavior both need review.
• Local focused test status: 1 focused test command blocked. The PR body says swift test --filter KimiProviderTests could not compile locally, leaving focused validation to CI or maintainer rerun.

Merge readiness
Overall: 🧂 unranked krab
Proof: 🧂 unranked krab
Patch quality: 🦐 gold shrimp
Result: blocked until real behavior proof is added.

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

Rank-up moves:

• [P1] Add redacted after-fix proof showing Kimi Monthly and Code 7-day quota rows populated from the new endpoint.
• Run swift test --filter KimiProviderTests plus the relevant widget snapshot tests in an environment with working Command Line Tools, or attach equivalent CI output.

Proof guidance:

• [P1] Needs real behavior proof before merge: No after-fix real behavior proof is attached; please add redacted live output, logs, terminal output, or a screenshot/recording showing the Kimi Monthly and Code 7-day rows, then update the PR body to trigger re-review or ask a maintainer to comment @clawsweeper re-review.

Risk before merge

• [P1] The new endpoint and ratelimitCode7d response shape are supported by the PR description and fixtures, but no attached after-fix live output verifies them against a real Kimi account.
• [P1] The contributor reports that focused Swift tests could not run locally, so CI or maintainer validation still needs to exercise the Kimi provider and widget snapshot tests.

Maintainer options:

1. Decide the mitigation before merge
   Land a narrow Kimi provider/widget fix after redacted proof confirms GetSubscriptionStats populates Monthly and Code 7-day rows, and focused Kimi/widget tests pass on a working toolchain.
2. Pause or close
   Do not merge this PR until maintainers decide whether the risk is worth taking.

Next step before merge

• [P1] This active PR is the implementation candidate; the remaining blocker is contributor-provided proof and focused validation, not a separate repair branch.

Security
Cleared: The diff changes Kimi parsing, same-origin endpoint selection, widget row propagation, tests, and changelog text; I found no concrete security or supply-chain regression.

Review details

Best possible solution:

Land a narrow Kimi provider/widget fix after redacted proof confirms GetSubscriptionStats populates Monthly and Code 7-day rows, and focused Kimi/widget tests pass on a working toolchain.

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

No high-confidence live reproduction is attached. Source inspection confirms current main still calls the endpoint the PR reports as returning 404 and only persists one Kimi extra widget row, but I did not run a real Kimi account probe.

Is this the best way to solve the issue?

Unclear until proof is added: the implementation is the narrow source-level fix for the reported endpoint drift and missing extra row propagation, but the live endpoint contract and rendered result still need after-fix evidence.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against b00797537dbf.

Label changes

Label changes:

• add P2: The PR targets a Kimi quota regression and a related quota-row improvement with limited provider/widget blast radius.
• add rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
• add status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: No after-fix real behavior proof is attached; please add redacted live output, logs, terminal output, or a screenshot/recording showing the Kimi Monthly and Code 7-day rows, then update the PR body to trigger re-review or ask a maintainer to comment @clawsweeper re-review.

Label justifications:

• P2: The PR targets a Kimi quota regression and a related quota-row improvement with limited provider/widget blast radius.
• rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: No after-fix real behavior proof is attached; please add redacted live output, logs, terminal output, or a screenshot/recording showing the Kimi Monthly and Code 7-day rows, then update the PR body to trigger re-review or ask a maintainer to comment @clawsweeper re-review.

Evidence reviewed

What I checked:

• AGENTS policy read: The full repository AGENTS.md was read; its guidance favors focused parser/provider tests and avoiding live credential-provider validation unless explicitly requested. (AGENTS.md:1, b00797537dbf)
• Current main still uses the old Kimi endpoint: KimiUsageFetcher.subscriptionStatURL points at MembershipService/GetSubscriptionStat on current main, matching the PR's reported obsolete endpoint. (Sources/CodexBarCore/Providers/Kimi/KimiUsageFetcher.swift:13, b00797537dbf)
• Current main only carries the monthly extra widget row: Kimi widget snapshot rows currently append only the kimi-monthly extra window, so a second Kimi extra quota row would not persist to widgets without this branch's propagation change. (Sources/CodexBar/UsageStore+WidgetSnapshot.swift:184, b00797537dbf)
• PR diff is narrow and targeted: The diff changes the Kimi endpoint, adds a Codable model for ratelimitCode7d, maps it to kimi-code-7d, updates Kimi widget row limits, and adds focused parser/widget tests. (Sources/CodexBarCore/Providers/Kimi/KimiUsageFetcher.swift:13, 6ef79606f5af)
• Validation and proof gap: The PR body reports git diff --check and swiftc -parse, but says swift test --filter KimiProviderTests could not run locally due an xcrun Command Line Tools failure; no after-fix live output, logs, screenshot, or recording is attached. (6ef79606f5af)
• Kimi subscription feature history: The merged monthly subscription work at [codex] Add Kimi monthly subscription usage #1839 added the membership subscription flow that this PR updates after upstream endpoint drift. (Sources/CodexBarCore/Providers/Kimi/KimiUsageFetcher.swift:13, 2fa61d07844c)

Likely related people:

• zhiyue: Authored the merged monthly subscription usage PR at [codex] Add Kimi monthly subscription usage #1839, which introduced the membership endpoint and monthly Kimi window now being updated. (role: introduced subscription behavior; confidence: high; commits: 2fa61d07844c; files: Sources/CodexBarCore/Providers/Kimi/KimiUsageFetcher.swift, Sources/CodexBarCore/Providers/Kimi/KimiUsageSnapshot.swift, Tests/CodexBarTests/KimiProviderTests.swift)
• joeVenner: Authored the merged Kimi widget support at Add Kimi widget support #1918, and blame attributes the Kimi widget row preservation and Kimi row limits to that merge. (role: recent widget owner; confidence: high; commits: 76a63dc55939; files: Sources/CodexBar/UsageStore+WidgetSnapshot.swift, Sources/CodexBarWidget/CodexBarWidgetViews.swift, Tests/CodexBarTests/CodexBarWidgetProviderTests.swift)
• steipete: Current-main blame attributes the Kimi subscription URL and monthly mapping lines to Peter Steinberger's release/mainline commit, and steipete has recent owner review activity on broader Kimi Code reporting work. (role: current Kimi source line owner; confidence: medium; commits: a3454640cf53; files: Sources/CodexBarCore/Providers/Kimi/KimiUsageFetcher.swift, Sources/CodexBarCore/Providers/Kimi/KimiUsageSnapshot.swift)
• Zihao-Qi: Authored the recently merged Kimi rate-limit ordering work at [codex] Fix Kimi rate limit ordering #1914, which shares the Kimi quota presentation invariants this PR extends. (role: recent Kimi presentation contributor; confidence: medium; commits: 94ed5d9ca522; files: Sources/CodexBarCore/Providers/Kimi/KimiUsageSnapshot.swift, Tests/CodexBarTests/KimiProviderTests.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.

[steipete]

Thanks @skyzer! An improved version landed in #1953 as dcccce8, preserving your authorship and changelog credit.

The original fork branch could not be updated because maintainer edits were disabled, so the repair had to land from a maintainer branch. Enabling “Allow edits by maintainers” on future PRs lets us improve the original branch directly.