What is the difference between 'warn' and 'block' actions?

The 'warn' action displays a message to the user but allows the AI to continue, while 'block' stops the AI entirely and returns a non-zero exit code to prevent the action.

Where are my custom hook rules stored?

Rules are stored as Markdown files in the .claude/ directory for project-specific settings or in ~/.claude/ for global rules that apply to all projects.

What is the Hookify skill for Claude Code?

Hookify is a specialized skill that allows you to define custom rules or 'hooks' that monitor Claude's actions to ensure safety, security, and adherence to coding standards.

Does this skill support regular expressions?

Yes, Hookify uses regex patterns to match file paths, shell commands, user prompts, and code content for highly precise rule enforcement.

Can I prevent Claude from deleting specific files or directories?

Yes, you can create a 'block' rule that triggers on specific bash commands or file patterns, effectively stopping Claude from executing destructive operations.

Claude Hookify: Custom Guardrails

Name: Claude Hookify: Custom Guardrails
Author: thkt

bythkt

•

Security & Testing

Enforce custom safety rules and coding standards in Claude Code using declarative hooks and automated triggers.

The Hookify skill empowers developers to create sophisticated guardrails that govern Claude's behavior throughout the development lifecycle. By defining declarative YAML rules, users can monitor file edits, bash commands, and user prompts to prevent dangerous operations like destructive shell commands or the accidental inclusion of sensitive data. Whether you need to block specific anti-patterns, enforce organizational coding standards, or require verification steps before task completion, this skill provides a robust framework for project-specific and global governance within the Claude Code environment.

Key Features

01Declarative YAML-based rule configuration for project or global scope

02Dual-action responses to either warn the user or block execution entirely

03Built-in CLI commands to list, configure, and generate new hooks on the fly

04Multi-event triggers for bash commands, file edits, prompts, and completion states

05Advanced regex pattern matching for granular control over commands and code

063 GitHub stars

Use Cases

01Automatically blocking the commitment of sensitive files like .env or debug logs to the repository

02Enforcing coding standards by detecting and warning against specific code anti-patterns or missing documentation

03Preventing the execution of destructive commands like 'rm -rf' or unauthorized 'chmod' changes

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thkt/claude-config creating-hooks

For use in Claude.ai and ChatGPT

Download Skill