agent-self-care（Agent 人自我照顾）

Agent Self-Care

Autonomous health monitoring and optimization for OpenClaw agents. Runs on cron or triggers manually.

Workflow

1. Check Sub-Agents

subagents action=list

Kill any stale sub-agents:

• Running >30 min without progress
• In "waiting" state
• Failed/errored

2. Check Processes

process action=list

Kill hanging processes:

• Running >10 min in background
• No output in 5 min

3. Check Session Health

session_status

Metrics to watch:

• Context usage >80% → trigger compaction
• Tokens growing unbounded
• Session age >2 hours → suggest refresh

4. Run Optimization Script

Execute scripts/optimize.sh which:

• Clears completed cron job artifacts
• Rotates logs if >50MB
• Reports health metrics

5. BMAD Retrospective (every 10 runs)

After 10 executions, run:

• Score last task 1-10
• Identify gaps
• Document improvements in memory/daily/YYYY-MM-DD.md
• Feed learnings into next run

Cron Schedule

Recommended: Every 5 minutes for active agents.

{
  "name": "agent-self-care",
  "schedule": {"kind": "every", "everyMs": 300000},
  "payload": {"kind": "agentTurn", "message": "Run agent-self-care skill"},
  "sessionTarget": "isolated",
  "enabled": true
}

Output Format

Report after each run:

🔧 Self-Care Report
- Sub-agents: X active, Y killed
- Processes: X running, Y cleared
- Context: X% used
- Health: ✅ GOOD / ⚠️ WARNING
- Retrospective: SKIP / COMPLETE

Key Principles

1. Always clean - Never leave stalled sub-agents or processes
2. Proactive - Don't wait for user to ask
3. Document - Log issues and improvements
4. BMAD - Continuous self-evaluation every 10 runs
5. Fast - Complete in <30 seconds