What are prompt-based hooks in Claude Code?

Prompt-based hooks are LLM-driven decision-making scripts that use natural language reasoning to validate tool usage, check task completion, or evaluate context-aware conditions without writing code.

How can I persist environment variables between Claude sessions?

By using a SessionStart hook and appending export statements to the file path stored in the $CLAUDE_ENV_FILE environment variable.

Can I use bash scripts with Claude Code hooks?

Yes, command-based hooks allow you to execute bash scripts for deterministic checks, file system operations, and integration with external CLI tools.

How do I prevent Claude from running dangerous commands?

You can use a PreToolUse hook with a matcher for 'Bash' to intercept, analyze, and either approve or block commands before they are executed by the agent.

What is the difference between plugin hooks and settings hooks?

Plugin hooks are defined in a hooks.json file within a plugin directory for distribution, while settings hooks are defined directly in the user's .claude/settings.json file.

Claude Code Hook Development

Name: Claude Code Hook Development
Author: jasonsie

byjasonsie

0•

개발자 도구

Provides comprehensive guidance and implementation patterns for creating event-driven automation hooks in Claude Code plugins.

This skill empowers developers to extend Claude Code's capabilities through custom hooks that intercept and react to lifecycle events. Whether implementing safety validations with PreToolUse, verifying task completion with Stop hooks, or automating environment setup via SessionStart, it provides standardized patterns for both LLM-driven prompt hooks and deterministic bash-based command hooks. It is an essential resource for teams looking to enforce coding standards, prevent dangerous operations, and create seamless, context-aware development workflows that integrate Claude directly into their existing toolchains.

주요 기능

01Environment variable persistence across sessions via CLAUDE_ENV_FILE

02Lifecycle management for PreToolUse, PostToolUse, Stop, and SessionStart events

030 GitHub stars

04Support for advanced prompt-based hooks for context-aware reasoning and validation

05Command-based bash hooks for deterministic system checks and external tool integration

06Comprehensive security patterns for path safety and input validation in autonomous agents

사용 사례

01Automatically loading project-specific context and environment variables upon session start

02Enforcing quality gates by validating task completion criteria before Claude stops work

03Blocking dangerous bash commands or unauthorized file writes during autonomous sessions

주요 기능

01Environment variable persistence across sessions via CLAUDE_ENV_FILE

02Lifecycle management for PreToolUse, PostToolUse, Stop, and SessionStart events

030 GitHub stars

04Support for advanced prompt-based hooks for context-aware reasoning and validation

05Command-based bash hooks for deterministic system checks and external tool integration

06Comprehensive security patterns for path safety and input validation in autonomous agents

사용 사례

01Automatically loading project-specific context and environment variables upon session start

02Enforcing quality gates by validating task completion criteria before Claude stops work

03Blocking dangerous bash commands or unauthorized file writes during autonomous sessions