fix(review): reflect a live manual-review label in the public PR comment

[JSONbored]

Closes #8067.

Summary

• A gate-blocked owner/admin PR gets manual-review applied as a fallback hold (src/settings/agent-actions.ts's generic manualHoldReason fallback). That label is deliberately maintainer-only-removable (src/services/agent-action-executor.ts's live-label guard, the "feat(enrichment): leftover console.log / debugger analyzer #3472 split-brain" check) — correct, since the bot must never silently override a real human freeze.
• But derivePublicCommentMergeFacts's heldForReview flag only ever checked for a hard-guardrail-path hit. It never checked the live PR for the manual-review label itself, so once the ORIGINAL triggering blocker cleared on a later pass, the public comment happily said "✅ approve/merge recommended, no blockers" while the executor kept silently denying every merge/approve attempt because the (now-stale, but still live) label was still on the PR — with zero visible explanation anywhere on the PR.
• Confirmed live on PR fix(review): stop bare hotkey/coldkey mentions from false-positiving as secret leaks #7994 (fixing Fix bare-word hotkey/coldkey false positives in containsSecretLikeText + static-message redaction bug #7981): tripped missing_linked_issue early from a wrong closing-keyword phrasing ("Implements Fix bare-word hotkey/coldkey false positives in containsSecretLikeText + static-message redaction bug #7981" instead of "Closes Fix bare-word hotkey/coldkey false positives in containsSecretLikeText + static-message redaction bug #7981"), got manual-review applied as the owner-fallback hold, the wording was fixed and the blocker cleared — but the label stuck around, and the PR sat fully green (CI + codecov + bot-approved) for 3+ hours before being merged manually. Full root-cause trail (audit_events, executor logs) is in fix(review): stale manual-review label silently blocks merge with no visible explanation #8067.
• Fix: broaden heldForReview to also go true whenever the configured manual-review label (default "manual-review", honors a repo's manualReviewLabel override/disable) is present on the live PR — regardless of which hold path originally applied it. Visibility-only change: the comment now correctly says "manual review recommended" instead of a misleading "ready"; the actual "only a maintainer can remove this label" safety behavior is untouched.

Test plan

• npm run typecheck
• New tests in test/unit/processors-public-comment-merge-facts.test.ts: label present → held; case-insensitive match; label absent → unaffected; custom configured label name; manualReviewLabel: null disables the check entirely
• Full unsharded npm run test:coverage: 1085/1085 files, 20278 tests, 0 failures
• 100% line/branch coverage on every new/changed line (verified via targeted coverage run + the full suite's processors.ts aggregate: 97.36% stmts / 96.76% branch, no new lines in the uncovered set)

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[loopover-orb]

Tip

✅ LoopOver review result - approve/merge recommended

_{Review updated: 2026-07-22 18:37:08 UTC}

2 files · 1 AI reviewer · no blockers · readiness 100/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
This fix closes the real gap in #8067: `heldForReview` previously only checked the diff for a hard-guardrail path, never the live PR's manual-review label, so once the original triggering blocker cleared the public comment could say "approve/merge recommended" while the sticky manual-review label (agent-action-executor.ts's live-label guard) kept the executor from acting — a genuinely misleading, silent state confirmed live on PR #7994. The fix threads `prLabels` through `derivePublicCommentMergeFacts`, does a case-insensitive match against the configured (or default `AGENT_LABEL_NEEDS_REVIEW`) label, and correctly treats `manualReviewLabel: null` as an explicit opt-out. Test coverage is thorough, hitting label present/absent, case-insensitivity, custom label override, and the null-disable arm — all real branches introduced by this diff.

Nits — 4 non-blocking

• Verify `pr.labels` at the `maybePublishPrPublicSurface` call site (processors.ts:10892) is genuinely `string[]` at runtime and not raw GitHub label objects (`{name: string}[]`) — the internal `PullRequestFileRecord`/`LocalBranchAnalysisInput` convention suggests it's already normalized, but this diff itself doesn't show that type definition.
• The doc comment above `derivePublicCommentMergeFacts` (processors.ts:2011-2023) is quite long for a one-field addition; consider trimming once the fix(review): stop bare hotkey/coldkey mentions from false-positiving as secret leaks #7994 incident context is captured in the issue/PR history instead of inline.
• Double-check that every other caller of `derivePublicCommentMergeFacts` (if any exist beyond the one shown) was updated to pass `prLabels`, since the param is required and TypeScript would only catch this if strict — the diff shows just the one call site update.
• PR author also opened the linked issue — Link an issue that was opened by a different contributor, or provide a rationale for why this self-authored issue represents genuine discovery work.

Decision drivers

• ✅ Code review — No blockers (1 reviewer)
• ✅ Gate result — Passing (No configured blocker found.)

Context & advisory signals — never blocks the verdict

Signal	Result	Evidence
Linked issue	✅ Linked	#8067, #7981
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (2 linked issues).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 16 registered-repo PR(s), 14 merged, 290 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	JSONbored; Gittensor profile; 16 PR(s), 290 issue(s).
Improvement	✅ Minor	risk: clean · value: minor · LLM: moderate

Linked issue satisfaction

Addressed
The PR directly implements the issue's proposed fix by broadening heldForReview in derivePublicCommentMergeFacts to check for the live manual-review label (with case-insensitive matching and custom label override support), and adds targeted unit tests covering the exact scenario described in the issue.

Review context

• Author: JSONbored
• Role context: owner (maintainer lane)
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: Python, TypeScript, Ruby, Go, JavaScript, MDX, Shell, Solidity
• Official Gittensor activity: 16 PR(s), 290 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Start here: Treat this as maintainer-lane context rather than normal contributor-lane activity.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

🧪 Chat with LoopOver

Ask LoopOver a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @loopover ask <question> answers contribution-quality Q&A with source citations and freshness.
• @loopover chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @loopover mention with a real question is routed to the closest matching read-only command automatically — no exact syntax required.

Full command reference: https://loopover.ai/docs/loopover-commands

🧪 Experimental — new and may change.

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by LoopOver, a quiet PR intelligence layer for OSS maintainers.

• Re-run LoopOver review

[loopover-orb]

LoopOver approves — the gate is satisfied and CI is green.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.78%. Comparing base (25e2df1) to head (5cde64d).
⚠️ Report is 10 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8068      +/-   ##
==========================================
+ Coverage   91.82%   95.78%   +3.96%     
==========================================
  Files         734        1     -733     
  Lines       75019     3135   -71884     
  Branches    23037     1115   -21922     
==========================================
- Hits        68885     3003   -65882     
+ Misses       5034       45    -4989     
+ Partials     1100       87    -1013     

Flag	Coverage Δ
rees	?
shard-1	80.44% <100.00%> (+26.17%)	⬆️
shard-2	30.23% <66.66%> (-22.34%)	⬇️
shard-3	39.29% <100.00%> (-15.81%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/queue/processors.ts	95.78% <100.00%> (+<0.01%)	⬆️

... and 733 files with indirect coverage changes