Can I use hooks to block Claude from specific files?

Yes, you can use the PreToolUse hook to inspect the file path and exit with a code of 2 to block the agent from reading or writing to sensitive files.

Do hooks support automated code formatting?

Absolutely. By using a PostToolUse hook on 'Edit' or 'Write' events, you can trigger formatters like Prettier to run automatically whenever the agent modifies a file.

What are Claude Code hooks?

Hooks are shell commands that execute at specific points in the Claude Code lifecycle, allowing you to intercept, log, or block agent behaviors deterministically.

Where are hook configurations stored?

Hooks are defined in JSON settings files, which can be global (~/.claude/settings.json), project-specific, or local-only for private configurations.

What information do hooks receive during execution?

Hooks receive a JSON payload via stdin that includes the session ID, transcript path, tool name, tool input, and current working directory.

Claude Code Hooks

Name: Claude Code Hooks
Author: filipexyz

byfilipexyz

0•

Productivity & Workflow

Implements deterministic shell command hooks to automate and control Claude Code agent behavior throughout its lifecycle.

Claude Code Hooks enables developers to execute custom shell commands at specific lifecycle events within the Claude Code environment, providing unparalleled control over agent interactions. By configuring hooks for events such as tool execution, user prompts, and session starts, users can implement automated logging, enforce security policies by blocking sensitive file access, and trigger auto-formatting or notifications. This skill is essential for teams looking to standardize AI agent behavior, integrate Claude into existing workflows, and maintain granular oversight of automated code modifications through precise, event-driven triggers.

Key Features

010 GitHub stars

02Context-aware execution with detailed JSON data passed to every hook via stdin

03Trigger custom shell commands at 10+ distinct lifecycle events including PreToolUse and SessionStart

04Automate post-processing tasks like code formatting, linting, or external notifications

05Enforce deterministic control with the ability to block or allow agent actions based on logic

06Implement security gates to prevent the agent from accessing sensitive files or secrets

Use Cases

01Automatically running Prettier or ESLint after the agent finishes writing code to ensure style consistency

02Blocking the agent from modifying sensitive files such as .env, credentials, or production configs

03Logging all terminal commands executed by the agent to a central audit file for security compliance

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add filipexyz/plugins claude-code-hooks

For use in Claude.ai and ChatGPT

Download Skill