gen-agents-md

Generate pointer-based AGENTS.md files referencing existing repository files.

Updated Apr 29, 2024
One-click install
npx skills add https://github.com/takaokanbe/dotfiles --skill gen-agents-md
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gen-agents-md
Source: https://github.com/takaokanbe/dotfiles/tree/main/config/claude/skills/gen-agents-md
Command: npx skills add https://github.com/takaokanbe/dotfiles --skill gen-agents-md

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agent root documentation files like AGENTS.md and CLAUDE.md often become bloated with outdated, redundant explanations that decay over time, leading AI coding agents to ignore critical rules or follow stale guidance that no longer matches the repository state.

Core Features & Use Cases

  • Pointer-based doc generation: Creates or rewrites AGENTS.md to reference existing repository files instead of repeating explainable information, eliminating documentation decay.
  • Cross-agent compatibility: Migrates existing descriptive CLAUDE.md files to the standardized AGENTS.md + CLAUDE.md symlink structure, compatible with Claude Code and other AI coding agents.
  • Strict guardrails: Enforces line count limits (≤50 lines recommended, ≤200 lines maximum) and validates all referenced file paths exist to ensure docs remain usable for agents.
  • Use case: Use when setting up a new repository, updating an existing bloated agent doc over 200 lines, or migrating an old descriptive CLAUDE.md to the modern pointer-based format.

Quick Start

Use the gen-agents-md skill to generate a maintainable, pointer-based AGENTS.md for your current code repository.

Frequently Asked Questions about gen-agents-md

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent my CLAUDE.md agent documentation from becoming outdated and bloated?

Pointer-based agent documentation prevents CLAUDE.md bloat by referencing existing repository files instead of repeating explainable information, eliminating documentation decay. It enforces line count limits and validates file paths to keep AI coding agents guided by current rules.

What is the best way to migrate a descriptive CLAUDE.md to a cross-agent compatible AGENTS.md?

Migrating a descriptive CLAUDE.md to a cross-agent compatible AGENTS.md uses a symlink structure that references existing files. This pointer-based format replaces redundant explanations, enforcing strict line count limits and validating referenced paths to ensure long-term usability.

How do I set up an AGENTS.md file for a new repository using pointer design?

Setting up an AGENTS.md file using pointer design involves generating references to existing repository files rather than embedding descriptive explanations. This approach enforces line count limits under 200 lines and validates all referenced paths to ensure AI coding agents follow current conventions.

Does pointer-based agent documentation work with Claude Code and other AI coding agents?

Pointer-based agent documentation works with Claude Code and other AI coding agents by using a standardized AGENTS.md and CLAUDE.md symlink structure. This cross-agent compatibility ensures all referenced file paths are validated and line counts remain within usable limits.

When should I rewrite my existing AGENTS.md or CLAUDE.md repository setup?

You should rewrite your existing AGENTS.md or CLAUDE.md repository setup when the documentation exceeds 200 lines or contains outdated, redundant explanations. Rewriting to a pointer-based format references existing files and enforces strict line limits to restore AI agent usability.

Why does my AI coding agent ignore the rules in my bloated AGENTS.md file?

AI coding agents ignore rules in bloated AGENTS.md files because outdated, redundant explanations decay over time and no longer match the repository state. Transitioning to pointer-based documentation references existing files, enforcing non-derivable conventions within strict line limits to maintain agent attention.