Hooks
Summary
- Hooks are event-based extension points that let you run custom logic before, during, or after agent actions.
Key Points
- Typical hook timing includes pre-action, post-action, and error/failure handling.
- Hooks are shared as a concept across many AI tools, but exact configuration and event names are platform-specific.
- Common uses: logging, policy checks, formatting, notifications, and guardrails before risky operations.
- Keep hooks deterministic and fast; slow hooks degrade overall agent workflow.
- Use hooks for enforcement (rules), not heavy business logic when possible.
Examples
- Pre-command hook blocks destructive commands unless approved.
- Post-task hook writes a summary note to a changelog file.
- Error hook sends failure details to a monitoring channel.
- AI/General/Skills
- AI/General/Sub-agents
- AI/General/MCP