Short answer
AGENTS.md is a shared instruction file for coding agents. CLAUDE.md is mainly for Claude Code. Cursor rules are mainly for Cursor. Teams often use one shared source plus small tool-specific files.
Words used in this guide
- AIPM: a tool that installs and updates reusable AI instructions.
- Workflow: a set of steps used to finish a task.
- Repository or repo: a project folder whose changes are saved and tracked.
- Git: a tool that records file changes so people can review or undo them.
What this means
AGENTS.md is the shared layer
AGENTS.md gives coding agents a predictable place to read repo instructions. It is useful when a team wants one common file for agent behavior.
CLAUDE.md is for Claude Code
CLAUDE.md is useful when a repo has Claude Code-specific setup. It can include commands, project notes, and expectations that matter for Claude workflows. Since v2.1.277, Claude Code also reads AGENTS.md when a project has no CLAUDE.md, so a repo that only has AGENTS.md no longer needs one. See does Claude Code read AGENTS.md? for the exact rule and how to load both.
Cursor rules are for Cursor
Cursor rules guide Cursor inside the editor. They are useful for file patterns, coding style, review behavior, and project-specific rules.
Simple steps
- Put shared repo rules in one main instruction file.
- Add CLAUDE.md only for Claude-specific behavior.
- Add Cursor rules only for Cursor-specific behavior.
- Keep duplicated text short.
- Turn repeatable tasks into skills and install them with AIPM when several repos need them.
Where to go next
If you want to try this in a real project, start with how to use AIPM and the command reference. If you want to share your own workflow, read the publishing guide.
FAQ
Should I copy the same rules into every file?
No. The copies may stop matching each other. Keep one main file and add short notes only when a specific AI tool needs them.
Can AIPM install these files?
No. AIPM does not install AGENTS.md, CLAUDE.md, or Cursor rules; keep those in Git. AIPM installs skills: --target claude writes .claude/skills/<skill>/SKILL.md and --target codex writes .agents/skills/<skill>/SKILL.md, and Cursor loads skills from both folders.