Claude Code Hooks overview

Claude Code Hooks as a cross-cutting topic - the pre- and post-action interception points that allow automating responses to agent lifecycle events without modifying the agent's core behavior. Covers PreToolUse hooks for approval and input mutation, PostToolUse hooks for side-effects and logging, Stop hooks for driving orchestration loops, Notification hooks for user feedback, and hook configuration via settings.json. Enables the Babysitter harness pattern and automated CI/CD-style workflows within agent sessions.