EPS Dashboard

Overview / Motivation

Auto-filed by /daily 2026-07-16 from transcript mining (chat a996d49a, 20:56-20:58Z, #1090 follow-up session churn).

Goal

Make deliberate stop-then-respawn a one-shot operation.

Workflow gap

Bug observed: spawn_session.py stop --session-id leaves the crash-recovery registration + per-issue dispatch lease in place, so a deliberate stop->respawn gets the replacement killed as a duplicate (registration still names the stopped session, 690s into a 900s window) — the #1090 scope-broadening tonight took a manual unregister + two lease clears + 3 spawn attempts
Why it is a workflow gap: stop is the sanctioned manual path; leaving stale registration/lease state behind turns every deliberate restart into a guard fight.
Confidence (emitter): high (reproduced tonight: 3 spawn attempts + manual state clears)
verified-at-filing: grep -n 'dispatch.lease\|def.*stop' scripts/spawn_session.py -> lease machinery at L93-L237 (#843 M1) with no stop-side release call visible; incident: replacement spawn killed as duplicate at 20:57Z (transcript a996d49a)

Proposed change (candidate diff sketch — refine in planning)

on stop with the process confirmed dead, atomically unregister the crash-recovery registration and release the issue dispatch lease so an immediate deliberate stop->respawn is one-shot

Scope / surfaces

Primary target: scripts/spawn_session.py
Grep the workflow surface for the pattern before editing and update every hit; list them in the plan.

Constraints / invariants

scripts/workflow_lint.py --check-asks passes; ruff on touched files passes.
This session runs under a workflow_fix_target: Provenance line — it MUST NOT auto-route its own subagents' workflow-fix candidates (recursion guard).

Provenance

fingerprint: 7a1af9675caa