Can it help clean up an existing AGENTS.md file?

Yes, it can refactor current content by removing redundant information, merging similar rules, and ensuring alignment with your project's existing guidelines.

How do I trigger this skill in Claude Code?

You can trigger it by asking Claude to 'clean up AGENTS.md', 'update the agents file', or 'reflect today's learnings in AGENTS.md'.

Why does the skill always ask for permission before editing?

To ensure that only valid, project-wide rules are added and to prevent cluttering your documentation with temporary or overly specific implementation details.

What kind of content does this skill exclude?

It automatically filters out one-off implementation details, local optimizations for single files, and information already covered in existing documentation directories.

What is the main purpose of the agents-md skill?

It is designed to help you maintain a high-quality AGENTS.md file by converting session-specific learnings into generalizable, reusable project rules.

Agents.md Manager

Name: Agents.md Manager
Author: takashabe

bytakashabe

•

コンテンツ管理

Refactors and maintains AGENTS.md files to capture session learnings into reusable project rules and guidelines.

The agents-md skill is designed to transform the AGENTS.md file into a high-value repository of generalizable project rules and best practices. It helps developers extract key insights from coding sessions, filter out temporary implementation details, and maintain a concise set of architectural and workflow guidelines. By following a strict 'propose-agree-edit' workflow, it ensures that every addition to your project documentation is intentional, reviewed, and adds long-term value to the development lifecycle without creating redundancy.

主な機能

01Cross-referencing with existing guidelines to prevent duplication

02Intelligent filtering to exclude one-off implementation details

03Refactoring of existing AGENTS.md content to remove redundancy

04Strict 'propose-agree-edit' workflow for controlled updates

052 GitHub stars

06Automated extraction of session learnings into reusable rules

ユースケース

01Synchronizing session-derived best practices across a development team

02Converting a specific architectural decision into a permanent project rule

03Cleaning up and organizing a cluttered AGENTS.md file

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add takashabe/dotfiles agents-md

For use in Claude.ai and ChatGPT

Download Skill