Does it support different tech stacks?

Yes, the generator performs a repository analysis phase to detect languages, frameworks, and tools to tailor the documentation structure to your specific project needs.

What is an AGENTS.md file?

An AGENTS.md file is a specialized documentation file designed to provide AI coding assistants with high-level context, project patterns, and search hints while minimizing token consumption.

How does the 'nearest-wins' hierarchy work?

AI agents are instructed to prioritize the AGENTS.md file closest to the directory they are currently editing, ensuring they use relevant local patterns rather than outdated or generic global ones.

Can this skill handle monorepos?

Yes, it is specifically designed to analyze monorepo structures and generate unique, scoped documentation for different apps, services, and shared packages within the same repository.

How does this improve token efficiency?

By using 'JIT indexing' (search commands and globs) instead of full file content, it allows AI agents to find what they need without loading excessive data into their context window.

AGENTS.md Generator

Name: AGENTS.md Generator
Author: julianromli

byjulianromli

•

101

•

Learning & Documentation

Generates hierarchical, token-efficient documentation structures that optimize codebase context for AI coding assistants.

The AGENTS.md Generator helps developers organize their codebases for AI agents like Claude by creating a distributed documentation hierarchy that improves model performance. Instead of overwhelming AI assistants with massive files, it implements a 'nearest-wins' approach with lightweight root files and detailed sub-folder guidance. This skill automates the analysis of your tech stack and directory structure to produce actionable, JIT-indexed documentation that reduces token usage while significantly improving the accuracy and context-awareness of AI-generated code.

Key Features

01101 GitHub stars

02Automated repository analysis to detect tech stacks, build systems, and directory maps.

03Standardized 'Definition of Done' and 'Universal Conventions' sections for team alignment.

04Hierarchical documentation structure featuring a lightweight root and detailed sub-folder files.

05Specialized templates for Design Systems, Databases, API layers, and Testing suites.

06Token-optimized JIT indexing using specific search commands (ripgrep) and file patterns.

Use Cases

01Standardizing coding patterns and architectural constraints for AI-assisted development teams.

02Setting up a new monorepo and needing AI-friendly documentation for distinct services and packages.

03Optimizing an existing enterprise codebase to reduce token costs for AI coding assistants.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add julianromli/opencode-template agents-md-generator

For use in Claude.ai and ChatGPT

Key Features

01101 GitHub stars

02Automated repository analysis to detect tech stacks, build systems, and directory maps.

03Standardized 'Definition of Done' and 'Universal Conventions' sections for team alignment.

04Hierarchical documentation structure featuring a lightweight root and detailed sub-folder files.

05Specialized templates for Design Systems, Databases, API layers, and Testing suites.

06Token-optimized JIT indexing using specific search commands (ripgrep) and file patterns.

Use Cases

01Standardizing coding patterns and architectural constraints for AI-assisted development teams.

02Setting up a new monorepo and needing AI-friendly documentation for distinct services and packages.

03Optimizing an existing enterprise codebase to reduce token costs for AI coding assistants.