EPS Dashboard

Overview / Motivation

Auto-filed by the /daily 2026-07-02 backfill problem sweep (route 2 — behavior/logic change, independent review required). Emitting context: transcript mining of the 2026-07-02 sessions.

Goal

Maintain a known-red-on-main baseline ledger (failing tests + repo-wide ruff count at a main SHA) that the Step 9c test-verdict diffs against, so sessions stop re-deriving pre-existing-ness

Workflow gap

Bug observed: On 2026-07-02 at least 7 sessions independently burned a verification round proving red tests/lint were pre-existing on main (2072-2084 ruff errors, failing audit-claim and thread-caps tests)
Why it is a workflow gap: The Step 9c verdict has no shared baseline of what is already red on main, so every code-path session pays a pristine-main cross-check round for failures its diff never caused
Confidence (emitter): medium

Proposed change (candidate diff sketch — refine in planning)

+ scripts/step9c_baseline.py refresh  # writes .claude/cache/
+   step9c-baseline.json {main_sha, failing_tests[], ruff_count}
+ select_step9c_tests / SKILL.md Step 9c: verdict compares the run's
+   failures against the ledger; only NEW failures block; stale ledger
+   (main advanced >N commits) => refresh first

Scope / surfaces

Primary target: scripts/select_step9c_tests.py, .claude/skills/issue/SKILL.md
Grep the workflow surface for the pattern before editing and update every hit; list them in the plan.

Constraints / invariants

Workflow-surface only — never experiment code, configs/, or tasks/.
scripts/workflow_lint.py --check-asks passes; ruff on touched files passes.
This session runs under EPM_WORKFLOW_FIX_SESSION=1 and carries a workflow_fix_target: Provenance line — it MUST NOT auto-route any of its own subagents' workflow-fix candidates (recursion guard).

Provenance

workflow_fix_target: scripts/select_step9c_tests.py, .claude/skills/issue/SKILL.md
fingerprint: b1dcc59feb5b
related_task: #852
source: /daily 2026-07-02 backfill problem sweep (route 2)