Squad MINOR findings — PR #4321 (coord-lifecycle regression guard)

Consolidated squad MINOR findings for PR: https://github.com/spec-kitty/spec-kitty/pull/4321 (`[#3867] test(coord-lifecycle): regression guard for the coord lane-lifecycle surface-authority class`, reviewed at head `50b93d80779f2404fe22cb7e95a6809bf77ddb1c`).

This checklist references code that does not exist on `main` until that PR merges.

- [ ] PR body `## Blast radius` — the recorded Discovery command `git diff --name-only upstream/main..HEAD`, re-run live, lists 31 files while `Files:` lists 2 (two-dot diff against a `main` that has since advanced; the genuine change is exactly the 2 new test files, confirmed via merge-base) — use the merge-base form `git diff --name-only $(git merge-base origin/main HEAD)..HEAD` so the evidence cannot go stale as main moves.
- [ ] tests/lanes/test_lane_lifecycle_characterization.py:278 — the characterization pin asserts the misleading `ERROR: issue-matrix.md` signal text, which is issue #3867's failure mode 5 (the gate names `.md` although the artifact is `issue-matrix.json`, and hides the per-row rule behind a generic "unresolved entries" message) — a diagnostics bug #3867 itself flags as a split candidate, but no split ticket was filed and PR #4321 closes #3867, so the bug loses its tracking — file/work the split: the approve gate should name the real artifact (`issue-matrix.json`) and surface the failing row + concrete rule (e.g. `deferred-with-followup` requires a `#NNN` / `Follow-up:` reference in `evidence_ref`).
- [ ] tests/lanes/test_lane_lifecycle_characterization.py:64 — SonarCloud reports 3.8% duplication on new code on PR #4321 (advisory gate, limit 3%): the two new test files share fixture/parametrize boilerplate — extract a small shared helper, or accept and note it in the PR body; advisory only (`continue-on-error`), not blocking.

Start with PR #4321 and tests/lanes/test_lane_lifecycle_characterization.py at lines 64 and 278, since the referenced code is not on main until that PR merges. Run the lane-lifecycle characterization tests, then address each checklist item: keep the diff evidence current, preserve accurate issue-matrix diagnostics, and resolve or document the reported test duplication. Done means all three findings are checked off with the PR body and tests consistent.

Written by the indexing model from the issue text.