Add assess_reconcile_inventory_report verifier to block complete queue scan
and all-already-landed claims without final-page metadata, requested page size,
pages fetched, per-page counts, and no-next-page proof. Wire into
build_final_report and export from review_proofs.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Add assess_already_landed_report_wording to review_proofs: when the
already-landed gate fires, final reports and controller handoffs must
use the reconciliation wording (Oldest open PR requiring action,
Eligibility class: ALREADY_LANDED_RECONCILE_REQUIRED, Candidate head
SHA, Reviewed head SHA: none, Review worktree used: false) and must not
use the legacy fields (oldest/next eligible PR, Pinned reviewed head,
Scratch worktree used) or claim a reviewed head SHA. In
non-already-landed reports, a populated Pinned reviewed head or
Reviewed head SHA fails unless validation and diff review actually
passed this session. Fails closed. Tests cover the correct
reconciliation wording, each forbidden legacy field, missing required
fields, populated reviewed SHA, normal reviewed reports, blocked-infra
reports, and validation-failed reports.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
A reviewer workflow approved and attempted to merge PR #278 although its
head commit was already an ancestor of master, ending in a Gitea 405 and
manual reconciliation. Add the missing pre-mutation gate:
- assess_already_landed_review_gate classifies the pinned candidate
head against the fetched target branch: NORMAL_REVIEW_CANDIDATE,
ALREADY_LANDED_RECONCILE_REQUIRED, or GATE_NOT_PROVEN (fail closed).
Already-landed PRs block approval and the merge API, allow
request-changes only for a real blocker, and require a reconciliation
handoff (PR number/title, candidate head SHA, target branch + SHA,
ancestor proof, linked issue status, recommended action, capability
proof for close/comment mutations). Unchecked ancestry, invalid or
missing SHAs, and a head changed since pinning all fail closed.
Mergeability stays a separate gate.
- assess_already_landed_report_state rejects APPROVED / MERGED /
READY_TO_MERGE wording and missing ALREADY_LANDED_RECONCILE_REQUIRED
state when the session's gate fired, complementing the text-marker
rules from #327 which need the report to admit the landed state.
Tests cover already-landed, normal, non-mergeable, changed-head,
unchecked-ancestry, and invalid-SHA gate paths plus all report-state
verdicts.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Block file edits in PR validation worktrees, require separate diagnostic
scratch worktrees with explicit labeling, and reject contradictory
File edits by reviewer: none claims.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Approving a PR whose full test suite fails is now gated behind complete
baseline proof. New assess_full_suite_failure_approval_gate:
- Full-suite pass: approval allowed, no baseline required.
- Full-suite failure: default action is REQUEST_CHANGES; approval is
blocked unless the #325 baseline validation proof passes (branches/
worktree, pinned baseline/PR SHAs, matching failure signatures,
clean before/after) and the proof also carries the PR and baseline
full-suite commands, evidence that new/changed tests passed, and a
non-empty explanation of why the failures are unrelated to the PR.
- Unknown/unproven full-suite result blocks approval.
- Main-checkout baseline comparison is rejected.
- Vague "same as master" claims without proof are rejected via the
existing pre-existing-failure claim detection.
- Non-approve decisions are never blocked; the gate only reports the
request-changes default.
Tests cover matching baseline failures, mismatched signatures,
incomplete proof (missing commands, new-test evidence, explanation),
main-checkout baseline, unknown suite result, vague claims, and
full-suite pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Add assess_linked_issue_consistency to review_proofs: reviewer final
reports must report a "Linked issue status" that matches the issue(s)
the selected PR was live-verified to link this session. Without live
proof the field may only say "not verified in this session"; with
proof, every linked issue must be named, and any other "Issue #N"
mention anywhere in the report fails as stale leakage from a previous
PR (the PR #280 / Issue #260-vs-#275 incident). Missing field fails
closed. Tests cover correct linked issue, stale previous issue in both
the status field and diagnostics, missing proof, unverified wording,
multiple linked issues, and PR-number mentions not being mistaken for
issue mentions.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Separate official PR-head validation from post-edit diagnostic runs in
reviewer reports. Block treating diagnostic passes as official validation
and require contaminated sessions to report recovery-required status.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
sysadmin
