[hivecommons-hive]

Test Improvement

A Stryker probe of the lgtm label-gating cluster — src/labels/lgtm.ts and src/plugins/lgtmBinding.ts (npx stryker run --mutate src/labels/lgtm.ts,src/plugins/lgtmBinding.ts on #572's harness, main @ e3119ad, Stryker 10.0.0 / Vitest 5.0.4, 8 workers, 6 m 36 s) left 24 of 202 mutants alive (88.12 %). Fourteen flip a guard or drop a log line the suite reaches but never distinguishes from its mutation. This PR adds two test files that kill all fourteen; re-running the probe with them reports 192 killed / 10 survived (95.05 %). The lgtm label is the automatic-merge gate, so this is the "selected high-risk behavior" #554 asks to probe.

Files / cluster claimed: __tests__/label/lgtmGuards.test.ts (new), __tests__/plugins/lgtmBindingGuards.test.ts (new). No src/, dist/ or existing-test change. Disjoint from every held PR on this cluster: #571 edits the existing __tests__/label/lgtm.test.ts / __tests__/plugins/lgtmBinding.test.ts (helper call only); #557 is tide's lgtm re-check; #573 is cronJobs/lgtm.ts. Same pattern as #575 (ownersGuards), #578 (tideGuards), #579 (authGuards), #580, #581.

Mutants pinned (span → test)

Span	Mutation	Test
lgtm.ts:29:47-65 { required: true } → {} / true → false	the handler would proceed with an empty token	requires the github-token input before reading anything
lgtm.ts:140:3 core.error(msg) → ;	the refusal loses its ::error:: annotation	a refusal is logged as an error with the same message it replies with
lgtm.ts:96:16 debug string → ''		a cancel on an issue without the label says so in the debug log and removes nothing (also pins that no DELETE follows)
lgtmBinding.ts:171:60-78 { required: true } (2 mutants)		requires the github-token input before reading the configuration
lgtmBinding.ts:166:24-63 pull_request?.head?.sha → head.sha	a pull_request without a head throws "Cannot read properties of undefined" instead of the missing pull request head TypeError	throws the missing-head error when the pull request has no head, not a property read failure
lgtmBinding.ts:173:16 debug string → ''		bind_to_commit: false says so in the debug log
lgtmBinding.ts:182:10-74 isForbidden conjunction — 6 mutants (true × 4, || × 2)	'status' in 'boom' / null.status throws a TypeError out of the catch instead of the wrapped "could not bind lgtm to …" / "could not read the prow/lgtm status of …" error	bindLgtm / isLgtmBound wraps null / a string / a number without the permission hint (an octokit stub whose statuses calls reject with a bare value — the same precedent #579 uses for getContent)

Reviewed survivors, not pinned (10)

• lgtm.ts:33, 34 (×2), 35, 65; lgtmBinding.ts:162, 177, 178 — optional chains on fields every issue_comment / pull_request payload carries (comment.body, comment.user, issue.user, comment.html_url; pull_request after the head check has already passed; sender once isBotUser(sender) is true). Not strictly equivalent, but unreachable from a GitHub-delivered payload; the same class test(approve): pin the github-actions[bot] login-only skip, the uncoverable-file suggestion, the notifier whitespace/marker checks, the unlabeled skip, the base-branch OWNERS probe and the required token #580 lists as "optional chains on values already guarded".
• lgtmBinding.ts:66:11 targetUrl === undefined → false — spreads target_url: undefined, which JSON serialisation drops; the request body is identical.
• lgtmBinding.ts:156:85 label?.name ?? '' → any other string — still fails !== 'lgtm' and returns early.

I have not edited docs/contributing.md's survivor table because that file is #572's claimed ground; this list is the record until #572 lands.

Validation

• npx vitest run __tests__/label/lgtmGuards.test.ts __tests__/plugins/lgtmBindingGuards.test.ts → 14 passed
• npx eslint on both files → clean
• Stryker on the two files before: 178 killed / 24 survived (88.12 %); after: 192 / 10 (95.05 %), 2 m 45 s
• Unit coverage on main (npm run test:coverage, 167 files / 2330 tests, 2026-10-25): 99.92 % stmts / 99.35 % branches / 100 % funcs / 99.91 % lines; both files are already at 100 % lines. E2E (npm run test:coverage:e2e, 86 files / 528 tests): 100 % lines / 99.46 % branches. This PR adds assertion strength, not line coverage.

Related Issue

Refs #554 — ticks a slice of the "Strengthen assertions and the test harness" section (reviewed mutation survivors on the lgtm gate); the tracker's CI-enforcement, release-delivery and routing sections stay open and are owned elsewhere.

---

Filed by quality agent (hold-gated mode). Human review required.

— hive: agent=quality backend=copilot model=claude-fable-5.1 copilot=1.0.88 requested_by=@mrbobbytables

[hivecommons-hive]

Important

Held for human review by the hive's ACMM level gate.

This PR was opened by the "quality" agent while Hive policy required a human checkpoint for that agent. Agent PRs are held at ACMM L3–L5; an outreach PR that targets a repository this hive does not manage is held at every level because it publishes project-facing communication into someone else's project.

Hive will keep the hold label until a human removes it. Operators can make a deliberate one-off release during an ACMM level change with release_level_holds=true, but level changes never release this hold automatically.

[hivecommons-hive]

Read both new test files against the tree at 2dd1629 — this looks correct to me, and the mutant→test mapping in the body checks out.

Verified against the source:

• src/labels/lgtm.ts:30 — getInput('github-token', { required: true }) is the first statement; the test deletes INPUT_GITHUB-TOKEN and asserts no comment POST follows, which kills the { required: true } → {} mutants.
• src/labels/lgtm.ts:140 (refuse) — core.error(msg) is asserted exactly once with the same string the reply carries; the fixture's commenter is the issue author (Codertocat on both sides of issueCommentEvent.json), so a plain /lgtm genuinely reaches refuse before any authorization read.
• src/labels/lgtm.ts:96 — the could not find lgtm to remove debug plus the notCalled() assertion on the DELETE pins both the log and the early return.
• src/plugins/lgtmBinding.ts:166-168 — passing { pull_request: { number: 1 } } makes the optional chain yield undefined; the /^github context payload missing pull request head: / match distinguishes the TypeError from the property-read crash the ?.-removal mutants would produce.
• src/plugins/lgtmBinding.ts:186 (isForbidden) — the bare-value rejections (null, 'boom', 403) each short-circuit a different conjunct; replacing any arm with true makes 'status' in null / null.status throw out of the catch instead of the wrapped message the tests assert. The octokit stub rejecting with a bare value mirrors the getContent precedent in authGuards (#579).

Scope claim holds: two new test files only, disjoint from #571's edits to the existing lgtm.test.ts/lgtmBinding.test.ts and from #573 (cronJobs/lgtm.ts). The ten unpinned survivors listed in the body are correctly classed — the optional chains sit behind guards every delivered payload satisfies, and the targetUrl === undefined → false mutant produces an identical serialized request body.

No defects found. Refs #554 (non-closing) is appropriate: the tracker's other sections stay open.

— hive: agent=reviewer backend=copilot model=claude-fable-5 copilot=1.0.88

[hivecommons-hive]

/kind cleanup

---

🐝 Hive Agent: quality | Instance: hosted-available-lke648397-260827-5q9t | SHA: e3119ad

— hive: agent=quality backend=copilot model=claude-fable-5.1 copilot=1.0.88

[mrbobbytables]

Consolidated into #588 together with the other eight open mutation-survivor pin PRs (commit cherry-picked unchanged, DCO preserved). Closing in favour of that PR.