Do I need to restart Claude to update rules?

No, rules are read dynamically. Any changes made to the `.local.md` files take effect immediately on the next tool execution or event trigger.

What are Hookify rules in Claude Code?

Hookify rules are markdown files with YAML frontmatter that define specific patterns for Claude Code to watch for. When a pattern matches a bash command, file edit, or prompt, Claude displays a custom message or blocks the action.

Can I prevent Claude from executing certain commands?

Yes, by setting the 'action' field to 'block' in the YAML frontmatter, you can stop Claude from executing specific bash commands or performing certain file edits that match your patterns.

Where are Hookify rule files stored?

Rules must be stored in the `.claude/` directory of your project root using the naming convention `hookify.{rule-name}.local.md`.

What regex syntax should I use for patterns?

Hookify rules use Python regex syntax. It is recommended to use unquoted patterns in YAML to avoid complex backslash escaping for common metacharacters like \s or \d.

Hookify Rule Architect

Name: Hookify Rule Architect
Author: ecielam

byecielam

0•

開発者ツール

Guides the creation and configuration of Hookify rules to enforce project-specific guardrails and automated feedback in Claude Code.

The Writing Hookify Rules skill provides a comprehensive framework for defining custom behavioral patterns within Claude Code. It enables developers to create markdown-based rules that monitor bash commands, file edits, and session completions in real-time. By leveraging YAML frontmatter and Python-based regex, users can implement automated warnings for dangerous operations, enforce strict coding standards, and ensure that vital checklists are followed before a task is finalized. This skill is essential for teams looking to bake project-specific tribal knowledge and safety protocols directly into their AI-driven development workflow.

主な機能

01Advanced multi-condition matching using regex, substring, and prefix/suffix operators

02Configurable actions to either warn the agent or block potentially destructive operations

03Standardized YAML frontmatter syntax for rule identification and event mapping

04Dynamic rule loading from the .claude/ directory for instant workflow updates

05Support for multiple trigger events including bash commands, file edits, stop signals, and user prompts

060 GitHub stars

ユースケース

01Automating the detection of debug code or sensitive information (API keys) before commits

02Preventing execution of dangerous shell commands like 'rm -rf' or privilege escalation

03Implementing mandatory completion checklists to verify tests and documentation before the agent finishes

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ecielam/claude-files writing-rules

For use in Claude.ai and ChatGPT