agentleFS
Mnorbert87Mnorbert87/claude-code-skills/skills/share-clean-package/SKILL.md

share-clean-package

Skill0 starsChanged 5 months agoMITView on GitHub

Turn a private/production project (bot, script, tool) into a clean, shareable package with no secret leakage. Use when the user says "send this to my friend", "I want to share this", "make a package out of this", and a running project must be prepared for a third party.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    AUDIT — mandatory: before packaging. grep the clean dir for:

    line 27

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill 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
SKILL.md

Turn a private project into a clean shareable package

When to use

The user wants to share one of their own projects (bot, script, tool) with someone else, and every piece of private data must be removed: API keys, wallet private keys, real addresses, trading/usage history, logs, env files, per-profile data.

Procedure

  1. New clean working dir /tmp/<project>-share/ — never work in the project's original location (needless contamination risk).
  2. Copy code only — cp the source files one by one (.py, .ts, .sh needed to run). Do NOT cp -r the whole folder.
  3. .env.example — copy the real .env with placeholders: every field empty, with comments on where to get the key (free API signup URLs). Never ship the real one.
  4. requirements.txt / package.json — only the minimal dependencies needed to run. Comment out optional packages (e.g. live-mode-only).
  5. README.md — three blocks minimum:
    • Strategy / what it does (include the underlying logic so the recipient understands why it works, not just a black box).
    • Install + run (recommend a safe/paper-mode default).
    • Pitfalls / critical notes (the knowledge you earned through mistakes — otherwise the recipient will make exactly the same ones).
  6. AUDIT — mandatory before packaging. grep the clean dir for:
    • Private-key patterns: 0x[a-fA-F0-9]{40,}|PRIVATE_KEY.*=.*\S
    • Real API keys: API_KEY.*=.*[A-Za-z0-9]{10,}
    • Concrete wallet addresses or user identifiers (know your own list of personal strings: emails, names, known wallet prefixes).
    • Result: every match must be an os.getenv / comment / code reference, never a hardcoded value.
  7. Package cd /tmp && zip -r <project>-share.zip <project>-share (or push to a fresh git repo).
  8. Deliver with an accompanying note that lists: what is included, the audit result (what it does NOT contain — more important than what it does), and the size.

Pitfalls

  • cp -r of the whole folder always fails — it drags in __pycache__, venv, node_modules, logs, history files, backups, and other projects living in the same folder (many tools share one parent dir). Copy file by file.
  • Audit BEFORE packaging — once it ships, it can't be un-shared. Secrets get cached and indexed even if you delete later.
  • The README is the most valuable part — code alone is 30 minutes of reverse engineering; with a good strategy explanation the recipient gets real knowledge.
  • Safe-mode default in the shared version (AUTO_TRADE=False or an equivalent flag). Leave the switch to live as a deliberate action by the recipient.
  • Don't include internal-infrastructure integration (dashboard endpoints, routes, app loaders). The recipient doesn't have your environment.

Verification

  • The audit grep runs clean (only variable references, no hardcoded secrets).
  • The package listing contains only the files meant to be shared.
  • The package size is proportional (10–100 KB for a small tool; MBs are suspicious — you probably dragged in a vendor dir).
  • The accompanying note explicitly states what is NOT included.

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.