Create Hooks FAQs

Question 1

What information is passed to a Claude Code hook?

Accepted Answer

Hooks receive detailed JSON input via stdin containing the event description, the specific tool being called (like 'Bash' or 'Edit'), and the tool-specific parameters (like file paths or shell commands). This data can be easily parsed using 'jq'.

Question 2

Can I target specific tools with my hooks?

Accepted Answer

Yes. Configuration in settings.json uses a 'matcher' field. You can set this to '*' to trigger for every tool call, or specify a specific tool name like 'Edit' or 'Bash' to ensure the hook only runs during relevant operations.

Question 3

What are Claude Code hooks?

Accepted Answer

Claude Code hooks are user-defined shell commands that execute automatically at specific points in the Claude Code lifecycle. They provide deterministic control over the environment, allowing you to trigger scripts based on events like tool usage or session starts.

Question 4

How can hooks improve my development security?

Accepted Answer

Hooks allow you to implement 'PreToolUse' validation. By using exit code 2 in a hook script, you can block Claude from performing unauthorized actions, such as modifying production environment variables, deleting critical directories, or accessing sensitive credentials.

Question 5

When should I use the PostToolUse event?

Accepted Answer

The PostToolUse event is ideal for automation tasks that should occur after code changes. Common use cases include running linters (like ESLint), code formatters (like Prettier or Black), executing unit tests, or creating automatic Git commits after a file is edited.

Create Hooks

Create Hooks

소개

주요 기능

사용 사례

소개

주요 기능

사용 사례