arc-skill-health-monitor
от trypto1019
DevOps & CloudКратко
Monitor deployed skills for performance drift, errors, and unexpected behavior changes.
Описание
Что делает этот навык
Monitors deployed skills for performance drift, errors, and unexpected behavior changes. Tracks latency percentiles, error rates, output format changes, and resource usage across invocations. Health data is stored locally per skill with configurable alerting thresholds.
Provides per-skill output drift detection that generic infrastructure monitors miss.
Когда его использовать
- Detecting latency increases after an upstream API update
- Catching error rate spikes in a production agent before users notice
- Identifying output format changes after a model behavior change
- Setting per-skill alert thresholds for automated notifications
- Reviewing 24-hour trends to spot memory or CPU accumulation
Пример сценария
User asks: check the health of my deployed summarizer skill
- 1
Runs health check against the summarizer skill command
- 2
Records execution latency, error count, and output snapshot
- 3
Compares current metrics against configured thresholds
- 4
Flags any output drift or threshold violations
Console report showing p50/p95 latency, error rate, and any threshold violations for the summarizer skill
Установка
clawhub install trypto1019/arc-skill-health-monitorВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.