📤

List your CLAUDE.md — one prompt does it

Turn the CLAUDE.md and instruction files you already have into listings by pasting one prompt into your AI. It collects your best files, strips secrets (API keys, client names), writes the descriptions, and submits — all automatically. Near-zero effort, about a minute.

1. 1

   Copy

   Copy the prompt with the button below.

2. 2

   Paste into your AI

   Paste it into Claude Code (or any agent) and send. ~1 minute.

3. 3

   Approve & done

   Your AI collects candidates and strips secrets, then asks per item 'OK to list?'. Approve the good ones — listing is automatic.

You are responsible for listing the strong AI instruction files (CLAUDE.md / skills /
dev playbooks) in this environment onto the marketplace "Makimono" (https://makimono-md.vercel.app).
Do the following carefully, in order.

[1. Collect] Find instruction files from:
  - ~/.claude/CLAUDE.md (user-global)
  - ~/.claude/skills/**/SKILL.md
  - CLAUDE.md and .claude/ playbooks/skills in each project
  List what you find (path + one-line purpose).

[2. Select] Keep only "generic know-how others can reuse as-is".
  Exclude: client/project-specific / personal notes / unfinished / too specific to this repo.

[3. Strip secrets (most important, non-negotiable)] From each candidate, remove or
  placeholder-ize:
  - Client / company / person names -> generalize or delete
  - API keys / tokens / passwords / DB connection strings / webhook URLs -> <YOUR_API_KEY> etc.
  - Internal IDs (Sheet ID / Drive folder ID / project ref / channel ID) -> delete or <ID>
  - Private internal circumstances or policies -> delete
  When unsure, do NOT delete silently — flag it and ask me (the human).

[4. Draft metadata] For each candidate produce:
  - title (what it accomplishes, ~60 chars)
  - summary (concretely what it does, 20+ chars)
  - category (pick one: Business Automation / Web Dev / Dev Process / Google Workspace /
    Data Collection / Social / Back Office / AI Behavior / Other)
  - scratchTokens (rough tokens if built from zero) / withMdTokens (with this MD; must be smaller)
  - price (0 = free for now) / author (your name or company)

[5. Approval gate] For each candidate, show me the sanitized full body + metadata and ask
  "OK to list this? (yes / edit / skip)". Only send the ones I approve. Never batch-send.

[6. Submit] POST approved candidates one by one:
    POST https://makimono-md.vercel.app/api/upload
    Content-Type: application/json
    body: {"title","summary","category","price","author","scratchTokens","withMdTokens","body"}
  Constraints: body 200-100,000 chars, summary 20+ chars.
  If you get 422 (secrets_detected), re-sanitize the flagged spot and resend.

[7. Report] Summarize results (title / category / submissionId / tokens saved).
  Note: submissions go live after Makimono's review.

Just share this URL — https://makimono-md.vercel.app/en/contribute — and anyone can grab the prompt.