Can I use hooks to block dangerous commands?

Yes, by using a PreToolUse hook with a command that returns exit code 2, you can block specific tool executions and provide feedback to the AI.

What are Claude Code hooks?

Hooks are specialized triggers that allow you to execute scripts or prompts at specific points in Claude's execution lifecycle, such as before or after a tool is used.

What is the difference between command and prompt hooks?

Command hooks run shell scripts or binaries, while prompt hooks send a text instruction to the LLM to evaluate whether an action should proceed.

Where are these hooks stored?

Hooks can be configured globally in ~/.claude/settings.json or specifically for a project in .claude/settings.json.

Does this skill support MCP tools?

Yes, you can target Model Context Protocol (MCP) tools using the 'mcp__server__tool' naming pattern in the hook matcher.

Claude Code Hooks Manager

Name: Claude Code Hooks Manager
Author: kkhys

bykkhys

•

生产力与工作流

Configures event-driven automation hooks to streamline workflows, enforce security, and automate tool-related tasks.

关于

The adding-hooks skill provides a comprehensive framework for implementing custom lifecycle events within Claude Code. It allows developers to automate repetitive tasks—such as auto-formatting code after edits, validating bash commands before execution, or injecting real-time context into prompts—by hooking into specific session events like PreToolUse and PostToolUse. By simplifying the setup of user-scoped and project-scoped configurations, this skill ensures consistent tool behavior across environments while enhancing security through automated pre-execution checks.

主要功能

Comprehensive event triggers including PreToolUse, PostToolUse, and SessionStart
Pattern-based tool matching using exact strings or regular expressions
Support for project-scoped, user-scoped, and plugin-bundled hook configurations
Dual hook types supporting both shell commands and LLM-based prompt evaluations
1 GitHub stars
Granular execution control using exit codes to allow or block tool actions

使用场景

Enforcing security policies by validating shell commands before execution
Injecting dynamic context or environment metadata at the start of a session
Automatically running linters or formatters after Claude modifies a source file

关于

主要功能

Comprehensive event triggers including PreToolUse, PostToolUse, and SessionStart
Pattern-based tool matching using exact strings or regular expressions
Support for project-scoped, user-scoped, and plugin-bundled hook configurations
Dual hook types supporting both shell commands and LLM-based prompt evaluations
1 GitHub stars
Granular execution control using exit codes to allow or block tool actions

使用场景

Enforcing security policies by validating shell commands before execution
Injecting dynamic context or environment metadata at the start of a session
Automatically running linters or formatters after Claude modifies a source file