---

name: monitoring description: How to monitor agent performance and system health tags: [monitoring, metrics, observability, debugging]

Monitoring Skill

Use this skill to monitor KOR agent performance and system health.

Commands

• /metrics - View current metrics summary
• /sessions - List active sessions

Tracked Metrics

• Agent lifecycle: starts, ends, errors
• Tool calls: invocations, latency, success rate
• Sessions: duration, message count

Viewing Metrics

Run the collector script directly:

Copypython3 ${KOR_PLUGIN_ROOT}/scripts/collector.py summary

Best Practices

1. Review daily: Check metrics summary to identify issues
2. Track errors: Low success rates indicate problems
3. Monitor latency: High tool duration may need optimization