Claude Code Hooks Mastery

Hooks let you run deterministic scripts at fixed points in Claude Code's lifecycle — before tool calls, after edits, on notifications, when the agent stops. This repo demonstrates every hook event with runnable examples: blocking dangerous commands, logging tool use, text-to-speech notifications, and chaining hooks into observability pipelines.

It's structured as a learning resource: each hook is a self-contained UV Python script with commentary, so you can lift exactly the pattern you need into your own .claude/settings.json.