source-control:babysit-prs: test_pr_queue_snapshot.py integration test doesn't assert the ignored bucket for Approve-with-nits classification

[kyle-sexton]

Lane: work · Category: improvement

Triggering example

claude[bot] PR review on #567 (fix for #499 — babysit snapshot Approve-with-nits classification) surfaced one minor test-coverage gap, explicitly assessed as non-blocking ("not a correctness hole... worth a follow-up but not a blocker. Ready to merge"), deferred here rather than fixed inline.

Observed vs expected

test_pr_queue_snapshot.py's ApproveWithNitsClassification.test_approve_with_only_nits_is_not_blocking asserts material == [] and blocking == [] for the end-to-end classification path, but doesn't assert that the item actually lands in feedback["ignored"]. The unit test in test_babysit_feedback.py does cover the ignored bucket, so there's no correctness hole today — but the integration test doesn't fully encode the invariant it's meant to demonstrate. Expected: add self.assertEqual(len(result["feedback"]["ignored"]), 1) to that integration test to tighten the end-to-end assertion.

Notes

Cosmetic/coverage-scoped only; no correctness, security, or architectural concern. Full bot review: PR #567 review comment (claude[bot]).

Work-class: C3 (bug-fix-shaped) — attended triage 2026-07-23, operator-ratified. 🤖

[kyle-sexton]

This was generated by AI during triage.

T1 — routing status: ready. Single-line test assertion addition (test_pr_queue_snapshot.py), already scoped exactly by the source bot review, no correctness gap today, no design decision.

[kyle-sexton]

Folded into the rolling paper-cut sweep #657 per the v11 intake throttle (paper-cuts batch as sweep lines, consequence findings stay standalone). Content preserved as a checklist line there; this closure is the fold, not a wontfix. [Control tower, triage-audit execution 2026-07-20; operator veto window]

[kyle-sexton]

This was generated by AI during triage.

Reactivated (triage lane — #657 sweep-conversion, rule 8h). This issue was administratively FOLDED-closed into #657 under the v11 intake throttle — not fixed: it still carries status: ready, and no PR closed it (closedByPullRequestsReferences: []). Restoring it to the ready frontier as its own dispatchable work-item, rather than filing a duplicate (its fully-specified content already lives here). Its #657 fold-line is checked off with a pointer back to this issue.