fix(mcp): word-boundary linked-issue closing keywords

[glorydavid03023]

Summary

extractLinkedIssues in packages/gittensory-mcp/lib/local-branch.js extracts GitHub closing-keyword issue references from the branch name, title, body, and commit messages to populate linkedIssues — which the MCP sends to the API for eligibility, score preview, and gate prediction. Its regex matched the closing keywords with no word boundary:

/(?:close[sd]?|fix(?:e[sd])?|resolve[sd]?|#)\s*#?(\d+)/gi

so a keyword embedded in a longer word matched and captured the trailing number:

extractLinkedIssues("hotfix 5")     // => [5]   (should be [])
extractLinkedIssues("prefixes 12")  // => [12]  (should be [])
extractLinkedIssues("unclosed 9")   // => [9]   (should be [])

A branch like hotfix-cache 5 or a commit postfix 3 would spuriously link an issue, feeding a false linked-issue signal into the local preflight.

Fix: anchor the keyword alternatives with \b, matching the two canonical server-side extractors that already do this — src/db/repositories.ts:5718 and src/signals/engine.ts:5241 both use /\b(?:close[sd]?|fix(?:e[sd])?|resolve[sd]?)…/. The MCP client-side extractor was the outlier that dropped the \b. The bare # branch stays boundary-free so #123 still matches anywhere. Pure helper — no schema or API change.

No issue because issue creation is restricted on this repo for outside accounts; this is a small, self-evident correctness fix.

Scope

• The PR title follows type(scope): short summary Conventional Commit format, for example fix(api): restore profile access checks.
• 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.
• I linked an issue, or this is small enough that the summary explains why an issue is not needed.

Validation

• git diff --check
• npm run actionlint
• npm run typecheck
• npm run test:coverage (see note below)
• npm run test:workers
• npm run build:mcp
• npm run test:mcp-pack (see note below)
• npm run ui:openapi:check
• npm run ui:lint
• npm run ui:typecheck
• npm run ui:build
• npm audit --audit-level=moderate
• New or changed behavior has unit/integration tests for new branches, fallback paths, and sanitizer boundaries

If any required check was skipped, explain why:

• The change is confined to the MCP package helper (packages/gittensory-mcp/lib/local-branch.js), which is plain JS outside Codecov's src/**/*.ts include, so it carries no patch-coverage obligation; a regression test was still added in test/unit/local-branch.test.ts (imported via the existing dynamic-import harness) and passes. build:mcp (the package's node --check) passes and typecheck is clean. I did not run the full test:coverage / UI / workers steps because they are unrelated to this MCP-only diff, and I could not run test:mcp-pack locally because this machine runs Node 24 while the repo pins Node 22 (.nvmrc); scripts/check-mcp-package.mjs throws ERR_INVALID_ARG_TYPE on Node 24 identically on unmodified main. CI runs the full validate matrix on the pinned Node.

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, low-noise, and does not imply compensation guarantees or optimization tactics.
• Auth, cookie, CORS, GitHub App, Cloudflare, or session changes include negative-path tests. — N/A: no auth/cookie/CORS/session changes.
• API/OpenAPI/MCP behavior is updated and tested where needed. (Behavior fix in the MCP local helper; covered by the added regression test. No request/response schema changed.)
• UI changes use live API data or real empty/error/loading states, not production mock/demo fallbacks. — N/A: no UI change.
• Visible UI changes include a UI Evidence section below. — N/A: no visible UI change.
• Public docs/changelogs are updated where needed; changelogs are only edited for release-prep PRs. (No docs/changelog change; MCP changelog is a release-prep artifact.)

Notes

• Regression test in test/unit/local-branch.test.ts asserts standalone keywords still link (fixes #5, Closes 12 and resolves #34, see #7, closes#3) while embedded-keyword substrings no longer do (hotfix 5, prefixes 12, unclosed 9).
• No UI Evidence section: there is no visible UI, frontend, docs, or extension change.

[loopover-orb]

Tip

🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩

✅ Gittensory review result - approve/merge recommended

_{Review updated: 2026-07-01 06:42:29 UTC}

2 files · 1 AI reviewer · no blockers · readiness 73/100 · CI green · clean

✅ Suggested Action - Approve/Merge

• safe to merge

Review summary
The change correctly narrows the MCP linked-issue extractor so closing keywords only match at a word boundary while preserving bare # references. The added regression test covers both the intended standalone keyword cases and the previously false-positive embedded-keyword cases. I do not see a reachable correctness break in the visible diff.

Nits — 5 non-blocking

• nit: packages/gittensory-mcp/lib/local-branch.js:563 now exports a helper only so the unit test can import it; if this package treats named exports as public API, prefer testing through the existing metadata collection path or document that this helper is intentionally exported.
• nit: test/unit/local-branch.test.ts:1742 covers helper behavior directly, but it would be stronger to include one end-to-end metadata case proving a branch name or commit message like `hotfix 5` no longer populates `linkedIssues`.
• packages/gittensory-mcp/lib/local-branch.js:563: either keep `extractLinkedIssues` explicitly public or move the regression through `collectLocalBranchMetadata` so the test exercises the same path the MCP uses.
• test/unit/local-branch.test.ts:1742: add one collection-level assertion for an embedded-keyword branch or commit message to guard the wiring, not just the regex helper.
• Readiness score is below the configured threshold — Use the readiness panel as advisory maintainer context; the score does not block this PR.

Signal	Result	Evidence
Code review	✅ No blockers	1 reviewer
Linked issue	✅ Linked	#5, #34
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	❌ 5/25	Preflight is holding this PR; address the blocker before review.
Contributor workload	✅ 10/10	Author activity: 197 registered-repo PR(s), 131 merged, 9 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	glorydavid03023; Gittensor profile; 197 PR(s), 9 issue(s).
Gate result	✅ Passing	No configured blocker found.

Review context

• Author: glorydavid03023
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository registration is not available in the local Gittensory cache.
• Public profile languages: JavaScript, Python, TypeScript, C++, Kotlin, MDX, Rust
• Official Gittensor activity: 197 PR(s), 9 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Fix the blocker.
• Triage stale or unlinked PRs.
• Refresh registry data or choose a registered active repo.

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