Can hooks block Claude from executing a command?

Yes, using the PreToolUse hook, you can intercept a tool call and return a 'deny' decision, effectively blocking dangerous or unauthorized operations before they occur.

What is the difference between Stop and SubagentStop hooks?

The Stop hook triggers when the main Claude Code agent finishes its task, while the SubagentStop hook triggers when a spawned subagent completes its specific assignment.

How can I pass data between different hooks?

You can persist state or environment variables by writing to the $CLAUDE_ENV_FILE during a SessionStart hook, making those variables available to all subsequent commands in the session.

What is the benefit of using prompt-based hooks?

Prompt-based hooks allow you to use natural language reasoning to evaluate tool calls or task results, making them ideal for complex, context-aware decisions that are difficult to script with traditional bash commands.

How do I make my plugin hooks portable?

Always use the ${CLAUDE_PLUGIN_ROOT} environment variable in your hook commands to ensure file paths are resolved correctly regardless of where the plugin is installed.

Claude Code Hook Development

Name: Claude Code Hook Development
Author: jawhnycooke

byjawhnycooke

生産性とワークフロー

Implements event-driven automation and validation patterns for Claude Code plugin development.

概要

The Hook Development skill provides a comprehensive framework for building sophisticated, event-driven extensions for Claude Code. It enables developers to intercept and react to lifecycle events—such as tool execution, session starts, and task completion—using either deterministic bash commands or flexible LLM-powered prompt hooks. By providing standardized implementation patterns for PreToolUse, Stop, and SessionStart events, this skill allows for robust security enforcement, automated context loading, and high-fidelity validation of AI-generated actions within any development environment.

主な機能

Secure implementation patterns for input and path validation
Prompt-based hooks for context-aware LLM validation
Comprehensive lifecycle event support (PreToolUse, Stop, SessionStart, etc.)
Environment variable persistence via CLAUDE_ENV_FILE
0 GitHub stars
Command hooks for deterministic bash-based automation

ユースケース

Enforcing security policies by blocking dangerous CLI commands or sensitive file access
Verifying task completion and code quality standards before the agent exits
Automating project context loading and environment setup at session start

概要

主な機能

Secure implementation patterns for input and path validation
Prompt-based hooks for context-aware LLM validation
Comprehensive lifecycle event support (PreToolUse, Stop, SessionStart, etc.)
Environment variable persistence via CLAUDE_ENV_FILE
0 GitHub stars
Command hooks for deterministic bash-based automation

ユースケース

Enforcing security policies by blocking dangerous CLI commands or sensitive file access
Verifying task completion and code quality standards before the agent exits
Automating project context loading and environment setup at session start