fix(review): scope refusal state to its board

[Chris0Jeky]

The Review queue now binds its access-refused state to the board whose request returned 403. Switching to another board clears the old authority claim before loading, so a 500 on the new board cannot inherit a false refusal. Permission-paused polling resumes for the new board.

An explicit second 403 retains repeated-refusal feedback across that scope change using request-local state. Same-board retry and generation guards remain intact; background reads do not acquire explicit-retry feedback.

Validation:

• 180 composable tests passed, including A403-to-B403 feedback and B403-to-C500 attribution regressions. Terra independently repeated all 180.
• Exact Chromium review-queue-recovery E2E passed 1/1, including the formerly failing feedback assertion.
• Typecheck, production build and scoped ESLint passed. The final docs-only main merge passed doc links (666 files).
• Original independent review and the narrow Terra CI-fix review found no remaining C/H blocker. Hosted CI must pass at ea6de1c before merge.

Refs #2214. The separately tracked refresh-health disclosure/counter attribution remains outside this correction. Evidence is preserved at C:/Taskdeck-wave/evidence/2214.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-09T09:29:54.264884Z	83f9920	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[Chris0Jeky]

Independent review at 13ad222: no confirmed HIGH or CRITICAL findings. Verified the B 403 -> C 500 handoff, generation/supersession guards, same-board retry behavior, and re-armed poll lifetime; the scope transition clears only the stale B authority claim. Focused spec: 179 passed; typecheck and scoped ESLint pass.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 13ad222047

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[Chris0Jeky]

Independent Terra review, round 1 — no confirmed Critical/High finding at head 869e592, live base 8d7cd7d (merge base 2fd93d6). The existing Codex P2 is correctly tracked on #2214 and is non-blocking under the bounded gate. Local evidence: npx vitest --run --maxWorkers=2 src/tests/composables/useReviewProposals.spec.ts (179 passed). No fix batch required.

[Chris0Jeky]

CI triage update (round 1 complete): E2E Smoke failed in review-queue-recovery.spec.ts:61. The Playwright artifact shows that, after an A -> B scope change and the second explicit intercepted 403, paper-review-access-revoked-retry is absent. At this head, loadProposalsWithOutcome clears the previous revoked state when the requested scope differs before it issues B's read; the next 403 is therefore treated as a first refusal and does not set retry feedback. The durable revoked-access panel remains visible. This is a confirmed MEDIUM feedback regression, not a flaky/unavailable-log conclusion; it is outside the C/H-only one-fix-batch rule. CI is red and this PR is parked pending a separately scoped repair/re-gate.

[Chris0Jeky]

The narrowly scoped CI recovery is implemented in 5d9bf9d and independently reviewed by Terra high with no C/H findings. It carries only a request-local prior-refusal fact into a new explicit 403; a 500 still clears the previous board claim and takes the ordinary failure path. The new A403-to-B403 regression and existing cross-scope 500 protection pass. Author: 180 composable tests, typecheck/build/scoped lint, exact Chromium review-queue-recovery E2E 1/1. Reviewer independently repeated 180/180 tests. Final head ea6de1c additionally merges current main with docs-only changes; doc links666 pass. Fresh hosted CI is pending. The separate refresh-health attribution issue stays tracked on #2214.

[Chris0Jeky]

Disposable recovery checkout removed normally after confirming clean tracked source and existing PR ref equal to the final head. All owned E2E servers stopped. Generated bin/obj, dist, dependencies and the fresh synthetic E2E database were disposable; logs, review diff and cleanup inventory survive at C:/Taskdeck-wave/evidence/2214.

[Chris0Jeky]

Final current-base Terra reconciliation is clear. The refresh is a two-parent merge of the prior reviewed head and main f935ebc; its only added paths are the five representation-contract files, byte-identical to main, with no conflict-resolution edits. The added contract remains unregistered/unmapped with no runtime writer and no direct interaction with this reviewed seam was found. Original source/fix reviews remain applicable; fresh hosted CI at the refreshed head is the remaining gate. Closeout: PR #2814 and C:/Taskdeck-wave/evidence/HANDOFF.md.

[Chris0Jeky]

Fresh Terra base-interaction review completed against main 7ac99a7: no CRITICAL/HIGH blocker. Each current head was materialized and its merge tree reconciled cleanly. Original source/fix reviews remain applicable. For the control seams, control-trust tests passed 4/4 (#2792), nightly tests passed 35/35 with docs checks (#2791). Fresh current-head hosted CI remains required; this comment does not waive the maintainer-review gate on control paths.