claude-agents-md

Audit repository trees for AGENTS.md and CLAUDE.md wrapper configurations.

Updated Jan 5, 2026
One-click install
npx skills add https://github.com/d3ming/dot-agents --skill claude-agents-md
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-agents-md
Source: https://github.com/d3ming/dot-agents/tree/main/master/skills/claude-agents-md
Command: npx skills add https://github.com/d3ming/dot-agents --skill claude-agents-md

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Repositories that use AGENTS.md as the canonical instruction source often confuse Claude Code when frontend-specific files are missing, symlinked, or duplicated. This skill provides guidance and tooling to ensure CLAUDE.md wrappers import AGENTS.md correctly and that nested AGENTS.md files are covered.

Core Features & Use Cases

  • Standardize instruction wiring by recommending thin CLAUDE.md wrapper files that import AGENTS.md rather than duplicating or symlinking full instruction files.
  • Audit repository trees to find every AGENTS.md and verify the presence, type (real file vs symlink), and content of sibling CLAUDE.md files.
  • Use case: migrate a project to treat AGENTS.md as the single source of truth and generate a JSONL audit report that highlights missing or misconfigured CLAUDE.md wrappers.

Quick Start

Run the audit script to find every AGENTS.md and verify each has a real CLAUDE.md wrapper containing @AGENTS.md.

Frequently Asked Questions about claude-agents-md

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

FAQPage Schema
How do I keep AGENTS.md as the canonical instruction source for Claude Code?

To keep AGENTS.md canonical for Claude Code, create thin CLAUDE.md wrapper files that import AGENTS.md content rather than duplicating or symlinking full instruction files, ensuring Claude loads canonical instructions correctly.

How do I audit my repository for missing or misconfigured CLAUDE.md wrappers?

You can audit your repository by running a scan script that detects every AGENTS.md file, checks for sibling CLAUDE.md presence, verifies real file versus symlink status, and inspects content for @AGENTS.md imports, then emits a JSONL report.

Why should I use a CLAUDE.md wrapper instead of symlinking to AGENTS.md?

Using a thin CLAUDE.md wrapper that imports AGENTS.md is recommended over symlinking because symlinks can confuse Claude Code when loading instructions, whereas wrappers ensure canonical instructions are imported correctly without duplication.

Does the AGENTS.md audit script modify my repository instruction files?

No, the AGENTS.md audit script does not modify files. It only scans directories, detects CLAUDE.md presence and symlink status, inspects content for @AGENTS.md, and emits JSONL reports highlighting missing or misconfigured wrappers.

What is the best way to standardize instruction files across nested repository directories?

The best way to standardize instruction files across nested directories is to audit all AGENTS.md locations, verify each has a real CLAUDE.md wrapper containing @AGENTS.md, and ensure no symlinks or duplicated instruction files exist in the repository tree.

Can I use this audit for repositories with multiple nested AGENTS.md files?

Yes, the audit scans repository trees to find every nested AGENTS.md file and verifies the presence, type, and content of sibling CLAUDE.md files, generating a JSONL report that highlights any missing or misconfigured wrappers in nested locations.