daily-fix: watcher must not respawn duplicate per-issue sessions
kind: infra#wf-fix#wf-fix-fp:13a920921167#daily-auto-filed
track: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.
Overview / Motivation
Auto-filed by /daily 2026-06-29 problem sweep. Recurring wasted spawns across issue-664/665/739/740 + a mid-Phase-1.5 respawn race on #739.
Goal
The watcher never spawns a duplicate driver for an issue that already has a live (or actively-working) session.
Workflow gap
- Bug observed: watcher spawned duplicate per-issue sessions (each forced to detect-and-exit via the single-orchestrator guard); once respawned a 2nd #739 orchestrator mid-Phase-1.5 while its fact-checker subagent ran ~4 min, duplicating planning work.
- Why it is a workflow gap: the watcher respawn logic is workflow surface.
- Confidence (emitter): medium; multiple sessions. The guard catches it, but the spawns waste compute.
Proposed change
- Check the live-session registry before respawning a
/issue <N>driver; skip if a live session exists. - Treat a session with recent tool/subagent activity as alive (not stalled) before declaring it stalled and respawning.
Scope / surfaces
scripts/autonomous_session_watch.py.
Constraints / invariants
- Workflow surface only. Recursion guard: EPM_WORKFLOW_FIX_SESSION=1.
Provenance
- workflow_fix_target: scripts/autonomous_session_watch.py
- fingerprint: 13a920921167
Sessions: 716c2d95 (#739), 6bdade94 (#665), and #664/#740 duplicate-exit hits.
Activity