agentmd-maker

Generate repository-specific AGENTS.md and CLAUDE.md files from verified project evidence.

3|Updated Nov 30, 2025

Search Tags:#automation #documentation #validation #configuration #agents #project-management

Ready to run

npx skills add https://github.com/alpoxdev/hypercore-skills --skill agentmd-maker

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agentmd-maker
Source: https://github.com/alpoxdev/hypercore-skills/tree/main/skills/agentmd-maker
Command: npx skills add https://github.com/alpoxdev/hypercore-skills --skill agentmd-maker

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill solves the issue of inconsistent or hallucinated agent instructions by generating project-specific, verifiable AGENTS.md files based on actual repository evidence.

Core Features & Use Cases

Evidence-Based Generation: Automatically creates or refactors AGENTS.md and CLAUDE.md files by inspecting your project's manifests, CI workflows, and source structure.
Scoped Instructions: Supports nested AGENTS.md files for specific subtrees, ensuring rules are applied only where relevant.
Use Case: When onboarding a new AI agent to a complex monorepo, use this Skill to generate a precise AGENTS.md that contains only the actual build, test, and lint commands verified by your project's configuration files.

Quick Start

Use the agentmd-maker skill to analyze this repository and generate a grounded AGENTS.md file that includes only the verified build and test commands.

Frequently Asked Questions about agentmd-maker

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

FAQPage Schema

How do I create AGENTS.md files based on actual repository configuration?▼

Generate AGENTS.md files by inspecting repository manifests, CI workflows, and source structure to create grounded instruction files. This ensures build, test, and lint commands are verified against actual project configuration before being written.

Can I scope coding agent instructions to specific directories in a monorepo?▼

Scoped instructions support nested AGENTS.md files for specific subtrees, ensuring rules apply only where relevant. This allows distinct configuration files for different directories within complex monorepo structures.

How do I stop coding agents from using hallucinated build and test commands?▼

Prevent hallucinated commands by generating project-specific AGENTS.md files based on actual repository evidence. The system verifies commands against existing manifests and task definitions before writing them to configuration files.

Do I need repository read and edit permissions to generate project instructions?▼

Repository-level read, search, and edit capabilities are required to generate verifiable AGENTS.md files. These permissions ensure all generated commands and constraints are grounded in existing project manifests and task definitions.

What's the best way to refactor existing CLAUDE.md files for AI agent onboarding?▼

Refactor existing CLAUDE.md files by analyzing current manifests and CI workflows to update instructions with verified commands. This replaces outdated or hallucinated instructions with evidence-based project configuration.

When do I need nested AGENTS.md files instead of a single root configuration?▼

Use nested AGENTS.md files when onboarding AI agents to complex monorepos with varying subtree requirements. Scoped instruction files ensure rules are applied only where relevant within the repository structure.