Do hooks support MCP tools?

Yes, you can use matcher patterns like 'mcp__*' to target and intercept any Model Context Protocol tools used by the Claude Code agent.

What happens if a hook script fails?

If a hook script exits with a non-zero code other than 2, it typically results in a non-blocking error; an exit code of 2 explicitly blocks the operation and displays the stderr to Claude.

Can I block Claude from running certain commands?

Yes, by using the PreToolUse event and exiting your hook script with a specific exit code (2), you can block Claude from executing unauthorized or dangerous tools.

How do I configure these hooks?

Hooks are configured within the ~/.claude/settings.json or .claude/settings.local.json files using a standardized JSON structure defined in the skill documentation.

What are Claude Code Hooks?

They are custom scripts or commands that trigger during specific events in the Claude Code CLI lifecycle, such as before a tool runs or when a session ends.

Claude Code Hooks

Name: Claude Code Hooks
Author: 0xDarkMatter

by0xDarkMatter

•

Security & Testing

Implements a robust hook system to execute custom scripts before and after Claude Code tool invocations for enhanced control and automation.

About

Claude Code Hooks provides a programmable framework for intercepting and managing Claude Code operations through pre-defined events like PreToolUse, PostToolUse, and SessionStart. This skill allows developers to enforce security policies, automate audit logging, and validate tool inputs using custom Bash scripts or external commands. By integrating directly with the Claude Code CLI settings, it empowers users to create safer, more transparent, and highly customized AI-driven development workflows.

Key Features

Execution of custom shell scripts with JSON-based input/output for complex validation logic.
Direct integration with project-level or global settings for seamless deployment across environments.
Support for diverse lifecycle events including tool execution, permission requests, and session starts.
4 GitHub stars
Configurable timeout settings and error handling for blocking or non-blocking hook execution.
Flexible matcher patterns to target specific tools like Bash, Write, or MCP-based utilities.

Use Cases

Maintain detailed audit logs of all AI tool interactions for compliance and debugging purposes.
Trigger external notifications via Slack or email when Claude requests specific tool permissions.
Enforce security by blocking dangerous shell commands or validating file paths before execution.

About

Key Features

Execution of custom shell scripts with JSON-based input/output for complex validation logic.
Direct integration with project-level or global settings for seamless deployment across environments.
Support for diverse lifecycle events including tool execution, permission requests, and session starts.
4 GitHub stars
Configurable timeout settings and error handling for blocking or non-blocking hook execution.
Flexible matcher patterns to target specific tools like Bash, Write, or MCP-based utilities.

Use Cases

Maintain detailed audit logs of all AI tool interactions for compliance and debugging purposes.
Trigger external notifications via Slack or email when Claude requests specific tool permissions.
Enforce security by blocking dangerous shell commands or validating file paths before execution.