Step 5a/10d FAMILY_OF lint family: couple tests/test_step9c_*.py with the synced selector — family sync itself creates the WORKFLOW_INVARIANT hybrid-tree skew (issue-2342 gate round burned)
Highlight text on any card (body, plan, or an event) to anchor a comment, or leave a whole-task comment here. Mention @claude to summon a reply.
No comments yet.
Gap
The Step 5a family-atomic spec-freshness sync (and its Step 10d inline copy in .claude/skills/issue/steps/18-step-10d.md) puts scripts/select_step9c_tests.py, scripts/step9c_baseline.py-adjacent pins, tests/test_select_step9c_tests.py, and tests/step9c_workflow_invariant_manifest.txt in the lint family — but NOT the step9c-coupled test files themselves (tests/test_step9c_*.py, e.g. tests/test_step9c_base_identity.py, and any test file a main-side selector edit ADDS to WORKFLOW_INVARIANT, e.g. tests/test_step9c_constructed_path_consumers.py).
Incident (issue #2342 Step 10d, 2026-08-25)
The pre-gate Step 5a re-sync imported main's select_step9c_tests.py, whose WORKFLOW_INVARIANT literal names tests/test_step9c_constructed_path_consumers.py. That test file (added on main after the branch forked) and the updated tests/test_step9c_base_identity.py sit OUTSIDE the sync family, so the branch tree went hybrid: synced selector + stale/missing coupled tests. The Step 10d mapped-invariant gated leg then failed 3 NEW nodes (test_pinned_invariant_list_matches_live_tree on the missing file; 2 test_step9c_base_identity nodes on the stale copy) → verdict block → one full ~5h gate round burned + a manual 2-file git checkout origin/main -- repair (commit 2e4e566865 on issue-2342) + a second full gate round. The residual is documented in 18-step-10d.md as "(α) non-family rules-pin tests", but here the trigger was the family sync ITSELF creating the skew — the remedy is family-membership, not documentation.
Asked change
Add the step9c-coupled test files to the lint family in BOTH inlined FAMILY_OF copies (Step 5a in .claude/skills/issue/SKILL.md / its step file, and the Step 10d post-gate inline copy) and the SPECS/SPECS_10D pathspec lists: :(glob)tests/test_step9c_*.py (covers base_identity, baseline, constructed_path_consumers, and future additions), family lint. Rationale: any main-side WORKFLOW_INVARIANT/selector/step9c_baseline change is co-committed with its test files on main, so syncing the scripts without the tests is exactly the hybrid-tree shape the family-atomic design exists to prevent. Update the lint family drift pins (tests/test_issue_skill_lint_family_sync.py or equivalent) accordingly.
Provenance
Found by the /issue 2342 orchestrator at Step 10d (pre-push lint gate round 1 block, adjudicated as sync skew; repair commit 2e4e566865). Not a duplicate of #2374 (LESSONS.md coupling), #2423 (script-only sync satisfiability probe), or #2424 (sibling-script probe input set) — same class, distinct file coupling.
target_file: .claude/skills/issue/steps/18-step-10d.md wf_fix: true