What happens if both AGENTS.md and CLAUDE.md already exist?

The skill runs a conflict detection script to identify discrepancies in key sections and prompts you to choose which version to prioritize or if you want to perform a manual merge.

What specifically gets checked during the conflict detection phase?

It primarily checks critical sections like technology stacks, build commands, and project rules defined in the repository's conflict-rules.md file.

Can I recover my original CLAUDE.md if it is overwritten?

Yes, the skill automatically creates a timestamped backup file (e.g., CLAUDE.md.backup.2023...) before making any changes to your existing configurations.

How do I trigger the alignment process?

You can trigger it by asking Claude to 'align manifests', 'setup CLAUDE.md', or 'link agents' within your project directory.

Does this skill work on Windows systems?

The current implementation is optimized for Unix-based systems as it relies on symbolic link commands (ln -s) for file synchronization.

Agent Manifest Aligner

Name: Agent Manifest Aligner
Author: jo-minjun

byjo-minjun

•

Productivity & Workflow

Synchronizes and resolves conflicts between AGENTS.md and CLAUDE.md files to ensure consistent project instructions for AI agents.

The Agent Manifest Aligner is a specialized productivity skill designed to maintain harmony between core project configuration files. It automatically detects and links AGENTS.md to CLAUDE.md via symbolic links, ensuring that Claude and other AI agents always operate with a single source of truth for project rules. By identifying discrepancies in technology stacks or build commands and providing interactive resolution workflows, it prevents configuration drift and streamlines the setup of agent-ready repositories, especially during project initialization.

Key Features

01Automatic symbolic link creation between AGENTS.md and CLAUDE.md

02Automated timestamped backups for existing configuration files

03Intelligent conflict detection for tech stacks and build instructions

04Cross-file reference validation using specialized Python scripts

052 GitHub stars

06Interactive user-guided resolution for manifest discrepancies

Use Cases

01Initializing a new repository that requires standardized agent instructions

02Ensuring local development environments match global agent manifests

03Merging divergent project rules when CLAUDE.md and AGENTS.md both exist

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jo-minjun/claude-devkit agent-manifest-aligner

For use in Claude.ai and ChatGPT

Download Skill