feat: implement issue #1692 — [#1621 S2] Verify the addressed-marker against the pushed diff + honour maintainer dispositions

[don-petry]

User description

Closes #1692

Implemented by dev-lead agent. Please review.

---

CodeAnt-AI Description

Verify addressed review claims before resolving threads

What Changed

• Addressed replies must include one valid claim with a full commit ID and at least one changed file.
• Threads remain unresolved when the claim is missing, malformed, references a commit outside the pull request, or does not match a non-empty pushed diff.
• Maintainer requests are checked across the full thread, and required changes continue to block resolution until a later verified fix is pushed.
• Added guidance and automated coverage for claim validation, diff matching, maintainer dispositions, and fail-closed behavior.

Impact

✅ Fewer false review-thread resolutions
✅ Unresolved threads for fixes absent from the pushed diff
✅ Maintainer-required changes remain visible until addressed

💡 Usage Guide

Checking Your Pull Request

Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.

Talking to CodeAnt AI

Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:

@codeant-ai ask: Your question here

This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.

Example

@codeant-ai ask: Can you suggest a safer alternative to storing this secret?

Preserve Org Learnings with CodeAnt

You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:

@codeant-ai: Your feedback here

This helps CodeAnt AI learn and adapt to your team's coding style and standards.

Example

@codeant-ai: Do not flag unused imports.

Retrigger review

Ask CodeAnt AI to review the PR again, by typing:

@codeant-ai: review

Check Your Repository Health

To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[qodo-code-review]

ⓘ Qodo reviews are paused because your trial has ended. Ask your workspace admin to add credits to resume reviews. Manage billing

[codeant-ai]

🤖 CodeAnt AI — Review Status

Status	Commit	Started (UTC)	Finished (UTC)
✅ Incremental review completed	dc7dec9	Sep 08, 2026 · 11:37	11:37
✅ Incremental review completed	6faffbd	Sep 08, 2026 · 03:10	03:10
✅ Reviewed your PR	6f7474c	Sep 07, 2026 · 21:37	21:40

[codeant-ai]

---

Thanks for using CodeAnt! 🎉

We're free for open-source projects. if you're enjoying it, help us grow by sharing.

Share on X ·
Reddit ·
LinkedIn

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Advanced

Run ID: 2d7e04f0-1d7a-4aa1-a3cd-856d4a29c275

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gemini-code-assist]

Code Review

This pull request introduces a verifiable claim mechanism to ensure that review threads are only resolved when a machine-readable claim (containing a commit SHA and a list of files) is verified against the pushed diff. It updates the dev-lead prompts, modifies the review-fixing script, introduces a dedicated verification library, and adds comprehensive unit tests. The review feedback highlights a critical issue where an unguarded grep command under set -euo pipefail could prematurely terminate the script when no claim is found. Additionally, it suggests optimizing date comparisons by using UTC ISO-8601 timestamps and pure Bash lexicographical comparisons instead of spawning external jq processes inside loops.

[codeant-ai]

CodeAnt Nitpicks

2 code suggestions

1. A claim without the closing --> delimiter is still parsed when its JSON ends at the body, so malformed text can be accepted as a claim comment.

Possible bug · scripts/lib/addressed-claim-verify.sh:93

---

2. The version check accepts v:1.0 because jq treats it as equal to integer 1, although the documented schema requires an integer version.

Type error · scripts/lib/addressed-claim-verify.sh:100

[don-petry]

Dev-Lead — review-changes (applied)

Changes committed and pushed.

[don-petry]

Dev-Lead — fix-bot-comment (no-changes)

Agent reasoning

   - **CI Checks:** All are queued/pending (no failures, timeouts, or action_required)
   - **Review States:** Only COMMENTED reviews (no CHANGES_REQUESTED)
   - Result: Zero Tier 1 blockers
3. **Other Reviews:**
   - `gemini-code-assist[bot]`: COMMENTED state (observations, not blocking)
   - `codeant-ai[bot]` and `don-petry`: Empty COMMENTED reviews
**Conclusion:**
The SonarCloud comment is a neutral quality gate summary with no specific, actionable defects. There are no Tier 1 blockers (failed checks or change requests). Per the task guidance, a neutral overview describing passing metrics without identifying specific file/line issues is not actionable.
---
**No changes needed.** The PR has zero actionable issues from the bot comment and zero Tier 1 blockers. SonarCloud QG has passed cleanly. All other checks are in-flight (queued/pending).

[donpetry-bot]

Advisory bots were rate-limited; auto-approval is withheld until they recover. pr-review-sweep will re-review this PR after 2026-09-07T23:00:31Z.

[don-petry]

Dev-Lead — fix-bot-comment (no-changes)

Agent reasoning

Issues addressed: 0
- Quality Gate passed (no new defects found)
- Grep commands already properly guarded per existing code review
Files changed: None required
Skipped (informational): 1 (CodeRabbit review is COMMENTED, not CHANGES_REQUESTED; grep guards already present)
```
No changes needed — the PR is ready.

[donpetry-bot]

Advisory bots were rate-limited; auto-approval is withheld until they recover. pr-review-sweep will re-review this PR after 2026-09-07T23:21:27Z.

[don-petry]

Dev-Lead — fix-bot-comment (no-changes)

Agent reasoning

Issues addressed: 0
Files changed: none
Skipped (informational): 1
Summary:
- SonarCloud Quality Gate passed with 0 new issues and 0 security hotspots — no action required. This is a neutral, passing result, not an actionable defect.
- All 61 CI checks passed successfully.
- No CHANGES_REQUESTED reviews blocking the PR.
No bot findings require fixing. The PR is ready for merge.
```

[donpetry-bot]

Advisory bots were rate-limited; auto-approval is withheld until they recover. pr-review-sweep will re-review this PR after 2026-09-07T23:39:54Z.

[don-petry]

Dev-Lead — fix-bot-comment (no-changes)

Agent reasoning

- All review threads: ✅ Resolved
**Bot Comments:**
1. **SonarCloud comment:** Reports quality gate passed with zero issues—purely informational, no actionable findings
2. **Gemini-Code-Assist review:** Three findings, all already addressed:
   - ✅ **Grep guard fix** (PRRT_kwDOR9SdIs6gCA-Q): Resolved by wrapping grep with `|| true` in `acv_parse_claim` (scripts/lib/addressed-claim-verify.sh:85)
   - ✅ **Date format optimization** (PRRT_kwDOR9SdIs6gCA-a): Resolved by using Z-terminated UTC ISO-8601 format in `acv_gather_commit_facts` (scripts/lib/addressed-claim-verify.sh:276)
   - ✅ **Pure-Bash comparison** (PRRT_kwDOR9SdIs6gCA-e): Resolved by replacing jq date comparison with lexicographical string comparison in `resolve_addressed_bot_threads` (scripts/dev-lead-fix-reviews.sh:805)
All three gemini threads are marked resolved with proper fixes applied and addressed-marker replies posted.
---
**Result: No action needed** — all bot findings have been addressed and properly resolved. The PR is ready to merge.

[donpetry-bot]

Superseded by automated re-review at 6faffbd52c45561c750ff4f55ce813f22269900d — click to expand prior review.

Review — fix requested (cycle 1/3)

The automated review identified the following issues. Please address each one:

Findings to fix

Automated review — NEEDS HUMAN REVIEW

Risk: MEDIUM
Reviewed commit: 1150629f53257109b8191b2f06c4d0bf1e5f2c3d
Review mode: triage-approved (single reviewer)

Summary

Implements issue #1692 (epic #1621 S2): the dev-lead addressed-marker becomes a verifiable claim (<!-- dev-lead:claim {...} -->) checked against the pushed diff before a review thread is resolved, plus a full-thread scan honouring standing maintainer dispositions. The implementation is high quality and matches the pinned v1 schema in the issue body exactly, with fail-closed behavior throughout and thorough unit + harness-level regression coverage (including the PR #1044 shape). Escalating solely because 3 review threads were deliberately left unresolved by the PR author agent, one of which explicitly requests a maintainer decision on a design tradeoff.

Linked issue analysis

Closes #1692 — all acceptance criteria are substantively addressed:

• AC1 (verifiable claim): prompts (fix-reviews.md, fix-bot-comment.md, review-changes.md) emit the additive dev-lead:claim payload; the normative schema/parser lives in one place (scripts/lib/addressed-claim-verify.sh), per the pinned schema (v=1, full 40-char SHA, JSON files array parsed with jq — never IFS/cut/tr).
• AC2 (harness verifies): resolve_addressed_bot_threads() now gates on acv_parse_claim → commit-on-head → non-empty diff → file intersection (own diff first, cumulative <sha>^..HEAD second with a ::notice::), each failure logged and leaving the thread unresolved.
• AC3 (file-level hard gate, region advisory): path mismatch resolves but emits a ::warning:: naming both — matches the AC.
• AC4 (maintainer dispositions): comments(last:1) replaced with a full-thread fetch; acv_latest_maintainer_disposition scans ALL comments, reuses review_thread_is_agent_authored as the discriminator, requires the verified commit to strictly postdate the disposition, rc=2 fail-closed on unparseable timestamps.
• AC5 (fail closed): every parse/git/classification failure leaves the thread unresolved; pre-migration markers without a claim are treated as unverifiable (documented in the PR as the safe direction).
• AC6 (regression coverage): PR feat(canary-rings): onboard the 6 #482 reusables into the registry (#1036) #1044 shape (fix not touching claimed files), commit absent from head, malformed claim, and disposition-postdating cases all present in test_fix_reviews.bats; the pure helpers are exhaustively covered in test_addressed_claim_verify.bats.
• AC7: shellcheck and bats CI green; new suite registered in lint.yml.

Findings

Blocking (reason for escalation):

1. 3 unresolved review threads (codeant-ai, all 🟠 Major) with deliberate leave-open replies from the dev-lead agent. Two are inherent TOCTOU races around resolveReviewThread (no conditional mutation exists in GitHub's API; the pre-mutation fail-closed re-read is the best available mitigation — the replies are technically sound). The third, on scripts/lib/addressed-claim-verify.sh:270, flags that the cumulative <sha>^..HEAD fallback accepts any HEAD-ancestor, so a stale commit plus later unrelated changes to a claimed file could verify a false claim. The author agent explicitly wrote: "I'd rather surface this for a maintainer decision than reshape the verification contract inside a review-response pass." That is a direct request for human judgment on the verification contract — an auto-approver must not adjudicate it.
2. reviewDecision is REVIEW_REQUIRED with a pending review request to petry-projects/org-leads; mergeStateStatus is BLOCKED.

Non-blocking observations:

• acv_parse_claim accepts a claim whose JSON ends at the body without the closing --> delimiter (parameter expansion %% leaves the remainder when the suffix is absent). Fail-open only in the cosmetic sense — the JSON contract is still fully enforced. Worth a tightening pass alongside the maintainer decision above.
• A reply that quotes the claim prefix (e.g. in a code block) would count as a second claim → multiple-claims → thread stays open. Fails in the safe direction; noting for diagnosability.
• Triage classified this PR low-risk; I rate it MEDIUM — 871 additions to the org's merge-gate automation (size:XL) is a non-trivial logic change, though it strictly tightens (never loosens) the resolution gate.
• Secret scanning MCP tool unavailable in this run — relying on the green gitleaks check; no credentials or secret-like content in the diff.
• lint.yml change is a one-line bats-suite registration, consistent with the file's existing pattern (not a frozen caller stub).

CI status

All checks green: shellcheck, ShellCheck, bats, unit-tests, actionlint, CodeQL (actions + python), SonarCloud Quality Gate passed, gitleaks secret scan, agent-shield, holdout-guard, and all validate-* / permissions / stub-freeze gates SUCCESS. Ecosystem-specific audit jobs SKIPPED as expected. CodeRabbit reported SUCCESS (rate-limited earlier, since recovered).

---

Reviewed automatically by the PR-review agent (single-reviewer mode: fable 5). Reply if you need a human review.

Additional tasks

1. Resolve all unresolved review thread comments from other reviewers
2. Ensure all CI checks pass after your changes
3. Rebase on the target branch if behind
4. Do NOT modify files unrelated to the findings above

The review cascade will automatically re-review after new commits are pushed.

[donpetry-bot]

Advisory bots were rate-limited; auto-approval is withheld until they recover. pr-review-sweep will re-review this PR after 2026-09-08T04:18:47Z.

[don-petry]

Dev-Lead — fix-bot-comment (no-changes)

Agent reasoning

This is a neutral status summary with no specific, actionable defects tied to files or line numbers.
**CI checks:** All 52 checks completed successfully — no failures or Tier 1 blockers.
**Review state:** 
- `codeant-ai[bot]` — COMMENTED (empty)
- `don-petry` — COMMENTED (empty)
- `gemini-code-assist[bot]` — COMMENTED (not CHANGES_REQUESTED, so not a Tier 1 blocker)
## Conclusion
Per the guidelines: *"A neutral overview is not an actionable finding. If the bot comment merely describes or summarizes the diff without reporting a specific, actionable defect tied to a file/line, there is nothing to fix."*
The SonarCloud comment is a pass notification with no reported issues. No Tier 1 blockers exist, and no threads require replies or resolution.
**No changes needed** — the PR is clear.

[donpetry-bot]

Superseded by automated re-review at aa49f8fe20a6f64d74529c51eb8a8707faf73772 — click to expand prior review.

Review — fix requested (cycle 2/3)

The automated review identified the following issues. Please address each one:

Findings to fix

Automated review — NEEDS HUMAN REVIEW

Risk: MEDIUM
Reviewed commit: 6faffbd52c45561c750ff4f55ce813f22269900d
Review mode: triage-approved (single reviewer)

Summary

Implements issue #1692 (epic #1621 S2): the dev-lead addressed-marker becomes a verifiable claim checked against the pushed diff before a review thread is resolved, plus full-thread scanning of maintainer dispositions with fail-closed behavior. The implementation was reviewed in depth at 1150629f53257109b8191b2f06c4d0bf1e5f2c3d and found high quality; the only change since is a merge of main (PR #1694 files) that does not touch this PR's substance. The blocking findings from that review persist unchanged: 3 unresolved review threads, one of which explicitly asks for a maintainer decision on the verification contract.

Linked issue analysis

Closes #1692 — all acceptance criteria remain substantively addressed (unchanged since the prior review of 1150629f53257109b8191b2f06c4d0bf1e5f2c3d): additive dev-lead:claim v1 payload emitted by the three dev-lead prompts; normative parser/verifier centralized in scripts/lib/addressed-claim-verify.sh; resolve_addressed_bot_threads() gates on parse → commit-on-head → non-empty diff → claimed-file intersection (own diff first, cumulative range with ::notice:: fallback); full-thread maintainer-disposition scan requiring the verified commit to strictly postdate the disposition; fail-closed on every parse/git/classification failure; regression coverage for the PR #1044 shape, malformed claims, commits absent from head, and disposition postdating; new bats suite registered in lint.yml.

Findings

Blocking (carried forward — not addressed by the new commits, which are a merge of main only):

1. 3 unresolved review threads (codeant-ai, all 🟠 Major) remain unresolved at head:
   • scripts/dev-lead-fix-reviews.sh:702 and :815 — inherent TOCTOU races around resolveReviewThread (the leave-open replies are technically sound; no conditional mutation exists in GitHub's API).
   • scripts/lib/addressed-claim-verify.sh:270 — the cumulative <sha>^..HEAD fallback accepts any HEAD-ancestor, so a stale commit plus later unrelated changes to a claimed file could verify a false claim. The author agent explicitly deferred this to a maintainer decision on the verification contract. An auto-approver must not adjudicate it.
2. reviewDecision is REVIEW_REQUIRED with a pending review request to petry-projects/org-leads; mergeStateStatus is BLOCKED.

Non-blocking (carried forward): acv_parse_claim tolerates a missing closing --> delimiter (cosmetic fail-open; JSON contract still enforced); a quoted claim prefix in a reply counts as a second claim and safely keeps the thread open.

New issues since 1150629: none — the compare contains only the merge of main (PR #1694: persona PAT changes) with no modification to this PR's files beyond mechanical merge resolution.

Secret scan: run_secret_scanning MCP tool unavailable in this run; relying on the green gitleaks check. No credentials or secret-like content in the diff.

CI status

All checks green at 6faffbd52c45561c750ff4f55ce813f22269900d: shellcheck/ShellCheck, bats, unit-tests, unit, actionlint, CodeQL (actions + python), SonarCloud Quality Gate passed, Secret scan (gitleaks), agent-shield, holdout-guard, and all validate-*/permissions/stub-freeze gates SUCCESS. Ecosystem-specific dependency-audit jobs SKIPPED as expected. CodeRabbit SUCCESS.

---

Reviewed automatically by the PR-review agent (single-reviewer mode: fable 5). Reply if you need a human review.

Additional tasks

1. Resolve all unresolved review thread comments from other reviewers
2. Ensure all CI checks pass after your changes
3. Rebase on the target branch if behind
4. Do NOT modify files unrelated to the findings above

The review cascade will automatically re-review after new commits are pushed.

[don-petry]

Dev-Lead — fix-bot-comment (no-changes)

Agent reasoning

Issues addressed: 0
- Quality Gate passed with 0 new issues and 0 security hotspots — no actionable findings
Files changed: None
Skipped (informational): 0
```

[donpetry-bot]

Automated review — APPROVED ✓

Risk: MEDIUM
Reviewed commit: aa49f8fe20a6f64d74529c51eb8a8707faf73772
Review mode: triage-approved (single reviewer)

Summary

Implements issue #1692 (epic #1621 S2): the dev-lead addressed-marker becomes a verifiable claim checked against the pushed diff before a review thread is resolved, plus full-thread maintainer-disposition scanning with fail-closed behavior. The code was reviewed in depth in prior cycles at 1150629 and 6faffbd and found high quality; since 6faffbd the branch contains only merges of main (PRs #1689, #1709) touching none of this PR's files. Both blocking findings from the cycle-2 review are now cleared: all 7 review threads are resolved — the 3 previously-open codeant-ai Major threads (2 inherent TOCTOU races, 1 cumulative-fallback verification-contract question) were each explicitly resolved by the maintainer (don-petry) on 2026-09-08 with reasoned dispositions, providing exactly the maintainer decision the prior review required.

Linked issue analysis

Closes #1692 — all acceptance criteria substantively addressed (unchanged since the prior deep review): additive dev-lead:claim v1 payload emitted by the three dev-lead prompts; normative parser/verifier centralized in scripts/lib/addressed-claim-verify.sh; resolve_addressed_bot_threads() gates on parse → commit-on-head → non-empty diff → claimed-file intersection (own diff first, cumulative range with ::notice:: fallback; file-mismatch is advisory ::warning:: per AC3); full-thread maintainer-disposition scan requiring the verified commit to strictly postdate the disposition (AC4); fail-closed on every parse/git/classification failure (AC5); regression coverage for the PR #1044 shape, malformed claims, commits absent from head, and disposition postdating (AC6); new bats suite registered in lint.yml.

Findings

Resolved since cycle-2 review (no code changes needed — maintainer dispositions):

1. The 3 previously-blocking codeant-ai threads (dev-lead-fix-reviews.sh:702/:815 TOCTOU; addressed-claim-verify.sh:270 cumulative-fallback contract) are resolved by the maintainer with explicit 'Resolving as maintainer' rationale; codeant-ai had accepted each refutation via saved review instructions. The prior review's requirement that a maintainer — not an auto-approver — adjudicate the verification-contract question is satisfied.
2. All 4 gemini-code-assist threads resolved with verified fixes in earlier commits.

New issues since 6faffbd: none — the compare contains only merges of main (#1689 persona/schema changes, #1709 checkpoint-push) with no modification to this PR's 8 files.

Non-blocking (carried forward, cosmetic): acv_parse_claim tolerates a missing closing ' -->' delimiter (JSON contract still enforced); a quoted claim prefix in a reply counts as a second claim and safely keeps the thread open.

Security: lint.yml change is a single line registering the new bats suite — no workflow security smells. No secrets, eval, or injection patterns in the added shell code. run_secret_scanning MCP tool unavailable in this run; relying on the green gitleaks check.

Note: branch is BEHIND main (MERGEABLE); auto-rebase will bring it current after approval.

CI status

All code checks green at aa49f8f: shellcheck/ShellCheck, bats, unit-tests, unit, actionlint, CodeQL (actions + python), SonarCloud Quality Gate passed, Secret scan (gitleaks), agent-shield, Agent Security Scan, holdout-guard, and all validate-*/permissions/stub-freeze gates SUCCESS. Ecosystem dependency-audit jobs SKIPPED as expected. The three CANCELLED entries (dev-lead dispatch/ci-relay/resume) are agent-orchestration relay jobs superseded by concurrency, not code checks.

---

Reviewed automatically by the PR-review agent (single-reviewer mode: fable 5). Reply if you need a human review.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[donpetry-bot]

Advisory bots were rate-limited; auto-approval is withheld until they recover. pr-review-sweep will re-review this PR after 2026-09-08T13:06:53Z.

[donpetry-bot]

Automated review — APPROVED ✓

Risk: MEDIUM
Reviewed commit: a5f5e02072c3e1ad09f031ff15e4c229a9b58e23
Review mode: triage-approved (single reviewer)

Summary

Implements #1692: the dev-lead addressed-marker becomes a verifiable claim. New pure verifier lib (scripts/lib/addressed-claim-verify.sh) parses a machine-readable claim payload (schema v1: full 40-char SHA + non-empty JSON files array), verifies the commit is on the PR head with a non-empty diff intersecting the claimed files, scans ALL thread comments for standing maintainer dispositions (fixing the comments(last:1) blind spot from PR #1044), and fails closed on every ambiguity. Wired into resolve_addressed_bot_threads() with 235 lines of unit tests plus 4 harness-level wiring tests. Triage assessment confirmed: change strictly tightens the resolution gate; no new attack surface.

Linked issue analysis

Closes #1692 ([#1621 S2]). All six acceptance criteria are substantively addressed: AC1 claim schema emitted by prompts and documented normatively in the lib; AC2 harness verifies commit-on-head + non-empty diff + file intersection before resolving; AC3 file-level hard gate with advisory ::warning:: on thread-path mismatch (no false blocks); AC4 full-thread scan honours maintainer dispositions, resolving only when the verified commit postdates them; AC5 fail-closed throughout (unparseable claim/version/SHA/files, absent commit, empty diff, unorderable disposition all leave threads open); AC6 regression coverage includes the exact PR #1044 shape (fix not touching claimed files stays unresolved).

Findings

No blocking findings.

• Fail-closed design is consistently applied and well-tested: unknown schema version, abbreviated SHA, multiple claim comments, malformed JSON, commit absent from head, and empty diffs all leave the thread unresolved.
• Date-ordering logic is sound: commit date is forced to Z-terminated UTC ISO-8601 (TZ=UTC + format-local), matching GitHub createdAt width, so the lexicographic compare is chronologically correct.
• Advisory (non-blocking): the full-thread fetch uses comments(first:100) — a thread exceeding 100 comments would read a stale latest reply and could miss later comments. Extremely unlikely in practice; worth a follow-up if long threads ever occur.
• Advisory (non-blocking): the cumulative sha^..HEAD fallback means any later commit touching a claimed file verifies the claim, slightly weaker than per-commit verification. Deliberate (amended/split-commit case) and documented in both prompt and lib.
• Secret scan: the run_secret_scanning MCP tool is not available in this environment; noting per protocol. The gitleaks CI check is green, and the diff contains only placeholder example SHAs — no credentials.
• Prompt/doc changes keep the existing addressed-marker unchanged (additive claim), so review_reply_is_addressed_marker compatibility is preserved; pre-migration marker-only replies are correctly treated as unverifiable.

CI status

All validation checks green: shellcheck, bats/unit-tests, actionlint, CodeQL (actions+python), SonarCloud (Quality Gate passed, 0 new issues), gitleaks secret scan, agent-shield, lint, prompt-coverage, and all org policy gates SUCCESS. Skipped checks are ecosystem-conditional audits. The three CANCELLED entries (dev-lead dispatch/ci-relay/resume) are dev-lead orchestration jobs superseded by concurrency, not validation checks. 0 unresolved review threads; reviewDecision APPROVED.

---

Reviewed automatically by the PR-review agent (single-reviewer mode: fable 5). Reply if you need a human review.