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

Crash-fix relaunch flips a stale blocked status back to running when posting epm:run-launched, with a watcher reconcile flag for blocked tasks that have live healthy runs as backstop

Workflow gap

Bug observed: #742 ran healthy for ~11.5h while its task folder still said blocked (set during the June-30 crash streak, never flipped back); dashboard/watcher views stayed wrong until the user asked
Why it is a workflow gap: The crash-fix round recipe posts epm:run-launched but never reconciles a blocked status left by an earlier failed round, and no watcher pass flags a blocked task with a live run
Confidence (emitter): medium

Proposed change (candidate diff sketch — refine in planning)

SKILL.md crash-fix round step:
+ after a successful relaunch, if status == blocked: task.py set-status
+   <N> running --note 'relaunch succeeded; clearing stale blocked'
watcher: flag (never auto-flip) blocked tasks with fresh poll ticks +
  an alive workload pid, via the existing sidecar/Telegram channel

Scope / surfaces

Primary target: .claude/skills/issue/SKILL.md, scripts/autonomous_session_watch.py
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: .claude/skills/issue/SKILL.md, scripts/autonomous_session_watch.py
fingerprint: c15505ee3447
related_task: #742
source: /daily 2026-07-02 backfill problem sweep (route 2)