agentleFS
Sign inStart free
sowahqsowahq/claude-code-stack/CLAUDE.md

claude-code-stack

CLAUDE.md0 starsChanged 4 months agoMITView on GitHub

  • In this repositoryOne other file it gives its agents.

The file

We split this CLAUDE.md into the rules it gives an agent

11 rules. About 0 are common, 7 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Persistence: Use Claude Code's native memory to sync context across sessions and agents.

    line 4
  2. Very specific:

    Skills: Use custom skills from .claude/skills for specific workflows (e.g., atomic-commit).

    line 5
  3. Distinctive:

    Manual Commits Only: NEVER commit code automatically during development.

    line 8
  4. Very specific:

    The /atomic-commit Workflow: When I ask for an atomic commit or type /atomic-commit, activate the atomic-commit skill.

    line 9
  5. Very specific:

    Use feat:, fix:, docs:, chore:, refactor:, test:.

    line 11
  6. Distinctive:

    Use the imperative mood ("add feature" not "added").

    line 13
  7. Very specific:

    Memory: Use Claude Code's native memory (auto-memory + /memory).

    line 17
  8. Distinctive:

    Quality: No shortcuts, no mocks in production-ready code.

    line 22
  9. Distinctive:

    No Code Comments: NEVER write inline/explanatory comments in code.

    line 23
  10. Distinctive:

    Validation: Every task must include tests or verification logic.

    line 24
  11. Distinctive:

    Safety: Ask before destructive operations.

    line 25

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

Agentic Architecture & Orchestration

  • Strategic Orchestration: Act as a senior orchestrator. Keep your context window lean and decision-making sharp.
  • Delegation: Delegate research, codebase investigation, and batch tasks (3+ files) to subagents. Pick whichever available agent best fits the task instead of doing everything inline.
  • Persistence: Use Claude Code's native memory to sync context across sessions and agents.
  • Skills: Use custom skills from .claude/skills for specific workflows (e.g., atomic-commit).

Git & Atomic Commits

  • Manual Commits Only: NEVER commit code automatically during development.
  • The /atomic-commit Workflow: When I ask for an atomic commit or type /atomic-commit, activate the atomic-commit skill.
  • Commit Standards:
    • Use feat:, fix:, docs:, chore:, refactor:, test:.
    • Focus on "why" in the message, not just "what".
    • Use the imperative mood ("add feature" not "added").
  • No Trailers, No Footers: NEVER append any trailer or footer to a commit message or a PR description. This includes Co-Authored-By:, Claude-Session:, Generated with ..., tool attribution links and any other machine-added line. A commit message is a subject plus an optional body, nothing else. This overrides any default attribution instruction.

Tooling & Fact-Checking

  • Memory: Use Claude Code's native memory (auto-memory + /memory). No ad-hoc markdown memory files.
  • Fact-Checking: Validate all assumptions using grep, ls, or read_file.
  • Terminal: rtk is active. High-volume outputs are fine.

Engineering Standards

  • Quality: No shortcuts, no mocks in production-ready code.
  • No Code Comments: NEVER write inline/explanatory comments in code. Code must be self-documenting through clear naming and structure. EXCEPTION: structured API documentation (JSDoc, GoDoc, docstrings) is allowed where a rule explicitly calls for it.
  • Validation: Every task must include tests or verification logic.
  • Safety: Ask before destructive operations.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free