Diagram: in the root and tests/ folders, CLAUDE.md contains @AGENTS.md and imports the AGENTS.md next to it, the source of truth.

Shared Claude Code & OpenAI Codex instructions.

Store in AGENTS.md, reference from CLAUDE.md. AGENTS.md CLAUDE.md → @AGENTS.md /tests/ AGENTS.md CLAUDE.md → @AGENTS.md One caveat in the mechanics. Both merge nested instructions with the root file, but they discover them differently: • Claude Code discovers nested CLAUDE.md on demand when working with files in that directory. • Codex loads the AGENTS.md chain from the project root to the current working directory at session start. Official docs: Claude Code → https://code.claude.com/docs/en/memory#how-claude-md-files-load OpenAI Codex → https://learn.chatgpt.com/docs/agent-configuration/agents-md#how-codex-discovers-guidance