Can I block Claude from performing a specific action?

Yes, by using gatekeeper events like PreToolUse and returning an exit code of 2, you can veto an action and display a reason to Claude.

What is a hook in Claude Code?

A hook is an event-driven script or command that triggers before, during, or after Claude performs an action, allowing you to monitor, modify, or block its behavior.

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

A command hook executes a local shell script for logic, while a prompt hook uses Claude's LLM to evaluate a condition or generate context based on a prompt.

Where are hook configurations stored?

They are typically stored in .claude/settings.json for project-specific hooks or ~/.claude/settings.json for personal, global hooks.

Which programming languages are supported for hooks?

Hooks can be written in any language that can execute via the shell, including Bash scripts, Python programs, or simple inline terminal commands.

Hooks Builder

Name: Hooks Builder
Author: mike-coulbourn

bymike-coulbourn

•

보안 및 테스팅

Builds event-driven automation and security policies for Claude Code to monitor and control AI actions.

소개

Hooks Builder provides a comprehensive framework for creating event-driven automation within Claude Code, allowing developers to intercept and control Claude's behavior at every stage of the lifecycle. By defining hooks for events like tool execution, user prompts, or session starts, users can enforce security policies, automate code formatting, validate inputs, and ensure task completion. It supports blocking dangerous operations, modifying tool inputs, and logging activity, making it an essential tool for creating a secure and highly customized AI development environment.

주요 기능

Includes built-in security hardening checklists and testing workflows
Provides multi-language script templates for Bash and Python implementations
Allows input modification via the Transformer pattern for real-time prompt adjustments
Supports 10 distinct event triggers including PreToolUse, SessionStart, and UserPromptSubmit
Enables blocking (veto) capabilities for strict security and policy enforcement
3 GitHub stars

사용 사례

Blocking dangerous file operations or sensitive data access during AI sessions
Automating post-task code formatting or linting after tool execution
Injecting project-specific context or license headers into AI-generated code automatically

소개

주요 기능

Includes built-in security hardening checklists and testing workflows
Provides multi-language script templates for Bash and Python implementations
Allows input modification via the Transformer pattern for real-time prompt adjustments
Supports 10 distinct event triggers including PreToolUse, SessionStart, and UserPromptSubmit
Enables blocking (veto) capabilities for strict security and policy enforcement
3 GitHub stars

사용 사례

Blocking dangerous file operations or sensitive data access during AI sessions
Automating post-task code formatting or linting after tool execution
Injecting project-specific context or license headers into AI-generated code automatically