agent-failure-registry
por unleashedbelial
Coding Agents & IDEsResumen
Search the Agent Failure Registry for known agent failures, solutions, and lessons learned.
Descripción
Qué hace esta skill
A searchable database of documented agent failures, root causes, fixes, and lessons learned. Covers categories like API failures, auth expiry, rate limits, silent failures, and more. Also supports contributing new post-mortems via GitHub PR.
Surfaces structured, community-verified fixes for specific agent failure patterns faster than searching general forums or documentation.
Cuándo usarla
- Looking up known fixes when a Puppeteer automation script suddenly breaks
- Finding solutions before manually debugging an OpenAI rate limit error
- Submitting a post-mortem after resolving a tricky auth token expiry bug
- Browsing recent community submissions to anticipate common failure patterns
- Searching by error category to narrow down root causes across past incidents
Flujo de trabajo de ejemplo
User asks: 'Find known fixes for OpenAI API rate limiting'
- 1
Run search-registry.sh with --category rate_limit and --tag openai
- 2
Parse returned post-mortems for root cause and fix fields
- 3
Surface matching entries with summaries, applied fixes, and prevention steps
List of documented rate limit failures with confirmed fixes and prevention strategies
Requisitos
- GitHub account (only needed to submit new post-mortems via pull request)
Instalación
clawhub install unleashedbelial/agent-failure-registryEjecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.