Proposal: Handling Stale status=running Items in run_queue.json

Run: run_6f9a966b | Agent: agt_hq (Theo) | Date: 2026-02-27

Context

Tasks in run_queue.json may end up perpetually status=running due to system crashes, incomplete transitions, or process interruptions. A hygiene pass is needed to keep the queue aligned with SSOT.

Proposed Policy

1. Regular Audits — Weekly scan for items status=running longer than 24 hours.
2. Automatic Escalation — Flag stale running tasks and notify responsible personnel for manual verification.
3. Manual Verification — Validate each flagged task: genuinely in progress, or should be marked completed/failed/retry.
4. Status Update + Logging — Update status in run_queue.json with timestamps and user IDs for audit trail.
5. Failure Handling — Failed tasks trigger an investigation log for root cause analysis.

Data Fix Steps

1. Identify Stale Items — Script to list status=running items older than threshold.
2. Review and Update — Collaborate to review; update to done, failed, or queued as appropriate.
3. Documentation — Log findings and resolutions centrally.

Implementation

• Automate detection; keep manual override for status changes.
• No destructive deletions without approval.

RECEIPT runId: run_6f9a966b artifactPath: /Users/ENVOAI/.openclaw/workspace-theo/second-brain/brain/hq/agents/agt_hq/2026-02-27/run_6f9a966b.md