Where are hook configurations stored?

Hooks can be configured at the project level in .claude/hooks.json or at the user level in ~/.claude/hooks.json.

Can I block specific commands with hooks?

Yes, by using PreToolUse or UserPromptSubmit events with a 'block' decision, you can prevent Claude from executing potentially destructive or unwanted commands.

How do I debug my Claude Code hooks?

You should run Claude with the --debug flag, which reveals which hooks were matched, the execution status of commands, and the resulting outputs or errors.

What are Claude Code hooks?

Hooks are event-driven scripts or LLM prompts that execute in response to specific events in a Claude Code session, such as tool usage or session starts, allowing for custom automation and safety checks.

What is the difference between a command hook and a prompt hook?

Command hooks run standard shell scripts for tasks like logging or formatting, while prompt hooks use the LLM to perform reasoning-based validation or complex decision logic.

Claude Code Hooks Guide

Name: Claude Code Hooks Guide
Author: Eric-Ness

byEric-Ness

生産性とワークフロー

Automates and secures Claude Code sessions through event-driven shell commands and LLM-evaluated prompts.

概要

This skill provides comprehensive guidance for implementing Claude Code hooks, enabling developers to build project-specific automations, safety checks, and workflow customizations. By leveraging shell commands or LLM-evaluated prompts triggered by specific events—such as tool usage, session starts, or user submissions—users can programmatically control Claude's behavior to validate commands, inject context, or provide desktop notifications without modifying core code. It is an essential utility for developers looking to integrate Claude deeper into their local development environments and CI/CD mindsets.

主な機能

Blocking mechanisms to prevent unsafe or destructive actions
Event-driven automation for tool usage and session events
0 GitHub stars
Prompt-based LLM evaluation for complex decision logic
Granular tool matching using flexible regex patterns
Command-based shell script execution for logging and validation

ユースケース

Automatically formatting code with Prettier or Linters after edits
Preventing destructive Git commands or recursive file deletions
Sending desktop notifications when Claude requires user input

概要

主な機能

Blocking mechanisms to prevent unsafe or destructive actions
Event-driven automation for tool usage and session events
0 GitHub stars
Prompt-based LLM evaluation for complex decision logic
Granular tool matching using flexible regex patterns
Command-based shell script execution for logging and validation

ユースケース

Automatically formatting code with Prettier or Linters after edits
Preventing destructive Git commands or recursive file deletions
Sending desktop notifications when Claude requires user input