Claude Code CLAUDE.md Generator

Highlights

Example output

@AGENTS.md

# Claude Code Notes

This project shares its core agent instructions via AGENTS.md (imported above).
Add Claude-specific guidance below if needed; otherwise the imported content is sufficient.

Frequently asked questions

• Does Claude Code read AGENTS.md directly? ▾

  No. Claude Code only treats CLAUDE.md as memory. To use AGENTS.md alongside it, add a single @AGENTS.md import line at the top of CLAUDE.md. ConfigDeck inserts that import automatically when you select other tools too.
• Why is the @AGENTS.md import added automatically? ▾

  Cursor / Copilot / Codex treat AGENTS.md as their single source of truth, so Claude Code needs the import to follow the same rules. Without it, Claude Code ends up with a different ruleset and your tools' behavior diverges.
• How does the Claude Code-only mode work? ▾

  If you only use Claude Code, toggle 'Use Claude Code only' to write CLAUDE.md as the single source of truth without an @AGENTS.md import. AGENTS.md / Cursor MDC / Copilot instructions are excluded from the output.
• Where should I put CLAUDE.md? ▾

  CLAUDE.md at the project root is project memory; ~/.claude/CLAUDE.md is user memory. ConfigDeck outputs project memory at the root. You can keep your global personal memory in the user file separately to layer in your own working style.
• How do Skills fit in? ▾

  ConfigDeck generates eight .claude/skills/{name}/SKILL.md files together with CLAUDE.md. Claude Code auto-discovers the .claude/skills/ directory, so as soon as you unzip the output you can invoke commit, pr-create, debug and other workflows.