feat(review): surface the improvement signal in the PR panel

[JSONbored]

Summary

• Combines the deterministic structural-improvement tier (Aggregate deterministic structural-improvement sub-score #4742) and, where also active, the LLM
  tier's composed value judgment (Extend ModelReview with an ordinal improvement/value judgment #4743) into a new "Improvement" panel row — gated behind the
  improvementSignal converged feature (Config-as-code foundation: improvementSignal as a converged feature (global + per-repo) #4738); nothing renders anything when it resolves false for
  the repo.
• The band renders as static template copy (mirroring how "Readiness score: NN/100" already bypasses
  the public sanitizer safely), never as sanitizer-filtered dynamic text.
• Both the deterministic findings and the LLM rationale are re-checked against
  containsPrivatePublicTerm at render time — defense in depth, even though both are already
  claimed safe upstream.
• Wired into both buildPublicPrIntelligenceComment (legacy allRows) and
  buildPublicPrPanelSignalRows (unified-comment bridge) so the two never diverge, matching how the
  other seven rows are already hand-mirrored between them.
• New fields.improvementSignal toggle in the existing per-row show/hide mechanism, documented in
  .gittensory.yml.example alongside its siblings.
• Fixed a real pre-existing gap found while wiring the defense-in-depth check: containsPrivatePublicTerm's
  regex had "hotkey" but was missing its "coldkey" sibling.

Closes #4744
Part of epic #4737

Scope

• The PR title follows type(scope): short summary Conventional Commit format.
• This PR is focused and does not mix unrelated backend, UI, MCP, docs, dependency, and deploy changes.
• This follows CONTRIBUTING.md and does not reintroduce GitHub Pages, VitePress, site/, or CNAME.
• Linked issue: Closes PR panel: new advisory improvement-signal section #4744 (verified programmatically, see Notes).

Validation

• npm run typecheck — clean.
• npm audit --audit-level=moderate — 0 vulnerabilities.
• npx tsx scripts/check-engine-parity.ts — 22 hand-duplicated file pair(s) agree, no drift introduced.
• Full affected test suite (signals-coverage.test.ts, ai-review-advisory.test.ts,
  queue.test.ts, queue-2/3/4/5.test.ts) — 727/727 passing.
• Diff-scoped coverage cross-reference against coverage-final.json (not just the terminal
  summary): zero uncovered statements/branches on any added line in src/queue/processors.ts or
  src/signals/engine.ts. This included adding one new end-to-end test
  (queue-4.test.ts, "PR panel: new advisory improvement-signal section #4744: threads the improvement-signal row into the unified comment...")
  specifically to exercise the improvementSignalAllowed TRUE branch, since every other existing
  test in the suite leaves the feature off (the default) — that already covers the FALSE branch
  thousands of times over, but nothing previously drove the TRUE path at the processors.ts level.

Safety

• No secrets, wallet details, hotkeys, coldkeys, user PATs, private keys, raw trust scores, private
  rankings, or private maintainer evidence are exposed.
• Public GitHub text stays sanitized — a dedicated regression test asserts the new row never leaks
  forbidden vocabulary regardless of what findings/rationale feed it, mirroring this repo's existing
  public-safety invariant-test style.
• Auth/cookie/CORS/GitHub App/Cloudflare/session: N/A — no such surface touched.
• API/OpenAPI/MCP: N/A — no route/schema/tool changed.
• Public docs (.gittensory.yml.example) updated for the new fields.improvementSignal toggle.

UI Evidence

N/A — this is a Markdown/text panel-comment change, not a rendered UI surface; no screenshot tooling
applies. The exact rendered row text is shown verbatim in the new regression test's assertions and in
this PR's Validation notes above.

Notes

• Confirmed via gh pr view --json closingIssuesReferences that this PR shows exactly issue PR panel: new advisory improvement-signal section #4744 as
  its closing reference (not Epic: PR improvement signal — quality-delta scoring as the positive-axis counterpart to slop-risk #4737, the parent epic, which stays open — it has other sub-issues still
  pending).
• complexityDeltas/duplicationDeltas/patchCoverageDeltaPercent (three of the four deterministic
  inputs) have no live data source yet — REES's structured findings and a parsed Codecov number aren't
  threaded into the main app by any existing channel (see src/signals/improvement.ts's own header
  comment, from Aggregate deterministic structural-improvement sub-score #4742). Only the test-evidence axis can fire today. This is expected, not a gap in this
  PR's own scope — a future sub-issue's job.

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	gittensory-ui	0258ec5	Commit Preview URL Branch Preview URL	Jul 11 2026, 02:58 AM

[superagent-security]

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

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.25%. Comparing base (aa2bc7c) to head (0258ec5).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4954   +/-   ##
=======================================
  Coverage   94.24%   94.25%           
=======================================
  Files         445      445           
  Lines       38880    38900   +20     
  Branches    14160    14171   +11     
=======================================
+ Hits        36644    36664   +20     
  Misses       1577     1577           
  Partials      659      659           

Flag	Coverage Δ
shard-1	46.79% <60.86%> (+0.04%)	⬆️
shard-2	33.26% <30.43%> (-0.04%)	⬇️
shard-3	31.51% <13.04%> (+0.10%)	⬆️
shard-4	33.03% <86.95%> (+0.03%)	⬆️
shard-5	33.52% <30.43%> (+0.03%)	⬆️
shard-6	45.22% <34.78%> (-0.01%)	⬇️

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

Files with missing lines	Coverage Δ
packages/gittensory-engine/src/focus-manifest.ts	99.17% <100.00%> (ø)
src/config/gittensory-repo-focus-manifest.ts	100.00% <ø> (ø)
src/queue/processors.ts	95.62% <100.00%> (+<0.01%)	⬆️
src/signals/engine.ts	97.47% <100.00%> (+0.02%)	⬆️

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[loopover-orb]

Warning

🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨

⏸️ Gittensory review result - manual review recommended

_{Review updated: 2026-07-11 03:02:20 UTC}

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

⏸️ Suggested Action - Manual Review

• Touches a guarded path — held for manual review: This PR changes guardrail-protected path(s): .gittensory.yml (matched .gittensory.yml), src/queue/processors.ts (matched src/queue/**).

Review summary
This PR adds an advisory-only "Improvement" panel row combining the deterministic (#4742) and LLM (#4743) tiers, gated behind the improvementSignal converged feature, and threads it through both comment builders plus the field-toggle mechanism. The wiring (feature resolution, cache-hit/miss shape, defense-in-depth term filtering, focus-manifest field key) is consistent and well-tested, but the newly added queue-4.test.ts assertion contradicts the row's own stated invariant and is almost certainly the cause of the failing CI test shards.

Blockers

• test/unit/queue-4.test.ts (the `expect(postedBody).toContain("| Improvement | ⚠️ ℹ️ None detected | ...")` assertion near the end of the new PR panel: new advisory improvement-signal section #4744 test) expects a leading "⚠️" before the band label, but `IMPROVEMENT_BAND_LABELS` in src/signals/engine.ts:~4650 defines `none: "ℹ️ None detected"` with no ⚠️ prefix, and the surrounding doc comment explicitly states "Advisory icons only (✅/ℹ️, never ⚠️/❌)" — either this test's expected string is simply wrong (likely why validate-tests (3)/(5) and validate are failing in CI) or the unified-comment renderer is silently prefixing a warning icon for this new row, which would visually contradict the PR's own advisory-only design intent; trace which one it is and fix the actual mismatch rather than adjusting the assertion to match whatever currently renders.

Nits — 6 non-blocking

• src/signals/engine.ts: the extensive multi-paragraph doc comments on `buildImprovementSignalRow`/`improvementEvidenceText`/the exported args fields are quite long for this codebase's usual comment density — consider trimming to the non-obvious parts (the byte-identical/omission guarantee) and dropping the restatement of test assertions.
• `improvementEvidenceText`'s cap-at-2-findings behavior ("(+N more.)") is only exercised via `signals-coverage.test.ts`'s synthetic 4-finding fixture; consider also asserting the boundary case of exactly 2 findings (no "+N more" suffix) for full branch coverage.
• packages/gittensory-engine/src/focus-manifest.ts:438 and src/signals/engine.ts:4294/4684-4685 embed raw issue numbers (Aggregate deterministic structural-improvement sub-score #4742/Extend ModelReview with an ordinal improvement/value judgment #4743/PR panel: new advisory improvement-signal section #4744) inline in comments/arrays — harmless but the external tooling flags them as "magic numbers"; not worth changing.
• Before merging, run the failing shards (validate-tests (3) and (5)) locally against this exact assertion in test/unit/queue-4.test.ts to confirm whether the fix belongs in the test string or in whatever renderer prepends the icon.
• 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.
• Touches a guarded path — held for manual review — A maintainer must review and merge this change.

Signal	Result	Evidence
Code review	✅ No blockers	1 reviewer
Linked issue	✅ Linked	#4744
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 (1 linked issue).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 48 registered-repo PR(s), 40 merged, 312 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	JSONbored; Gittensor profile; 48 PR(s), 312 issue(s).
Gate result	⚠️ Not blocking	Advisory; not blocking this PR.

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: not available
• Official Gittensor activity: 48 PR(s), 312 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• 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.

_{🟩 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 Gittensory, a quiet PR intelligence layer for OSS maintainers.

• Re-run Gittensory review

[loopover-orb]

An AI reviewer flagged a likely defect, but its confidence was below this repository's configured close-confidence floor, so this is held for a maintainer to confirm instead of closing automatically. Resolve the flagged defect (see the review notes), or ask a maintainer to override.