Can I prevent Claude from running specific commands?

Yes, by setting the 'action' field to 'block' in the rule's YAML frontmatter, you can stop Claude from executing matched bash commands or performing specific file edits.

Does Hookify support complex matching logic?

Yes, Hookify supports an advanced format where you can define multiple conditions using various operators such as regex_match, contains, equals, starts_with, and ends_with.

Which events can trigger a Hookify rule?

Rules can be triggered by five event types: bash (command execution), file (edits and writes), stop (when the agent finishes), prompt (user input), and all (every event).

How do I create a new Hookify rule?

Create a file named .claude/hookify.{name}.local.md in your project root, add the required YAML frontmatter (name, enabled, event, pattern), and write the message you want Claude to see in the body.

What are Hookify rules in Claude Code?

Hookify rules are markdown files stored in the .claude/ directory that define specific patterns for Claude to watch. When a pattern matches a bash command or file edit, Claude displays a custom message or blocks the action.

Hookify Rule Architect

Name: Hookify Rule Architect
Author: FacuM

byFacuM

0•

开发者工具

Configures custom regex guardrails and automated checklists to monitor and guide Claude's behavior during development tasks.

The Writing Hookify Rules skill provides a standardized framework for creating project-specific guardrails within the Claude Code environment. By defining rules in local markdown files, developers can intercept dangerous bash commands, identify problematic code patterns like debug logs or security risks, and enforce mandatory checklists before a session concludes. This skill enables a 'YOLO' workflow with safety nets, allowing the agent to operate autonomously while adhering to domain-specific best practices and organizational safety standards through sophisticated regex matching and multi-condition logic.

主要功能

01Regex-based pattern matching for bash commands and file modifications

02Multi-condition logic support with operators like contains, starts_with, and equals

03Customizable warning and blocking actions to prevent unauthorized operations

04Automated completion checklists triggered on agent stop events

05Dynamic rule loading from local markdown files for immediate feedback

060 GitHub stars

使用场景

01Identifying and warning against production-unfriendly code such as console logs or hardcoded API keys

02Preventing the execution of destructive shell commands like 'rm -rf' or unauthorized 'sudo' usage

03Enforcing quality assurance checklists that Claude must review before completing a task

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add facum/yolo-agent writing-rules

For use in Claude.ai and ChatGPT

Download Skill