agentleFS
duncan-buildroomduncan-buildroom/profit-room-skills/skills/sanitize-project/SKILL.md

sanitize-project

Skill18 starsChanged 4 months agoMITView on GitHub

Sanitize and clean a project directory for public sharing. Use this skill whenever the user gives a file path and wants to package, share, or zip a project — even if they just say 'clean this up', 'prep this for sharing', 'sanitize this folder', 'remove my API keys', or 'package this project'. Also trigger when the user says /sanitize-project followed by a path. Handles removing build artifacts, credentials, secrets, caches, and anything that shouldn't be shared publicly.

  • Reads credentials
  • Deletes or force-pushes
  • Installs packages

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 7 are distinctive, and 7 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.

  • 7

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check that the directory exists before proceeding.

    line 20
  2. Very specific:

    All subsequent steps operate on $COPY_DIR, not the original.

    line 34
  3. Very specific:

    Use find + rm -rf to handle nested occurrences (e.g.

    line 74
  4. Distinctive:

    Search ALL remaining files for patterns that suggest credentials.

    line 111
  5. Distinctive:

    Use grep to scan:

    line 125
  6. Distinctive:

    Only warn/redact actual values, not placeholders

    line 136
  7. Very specific:

    Check if it contains real values (not just placeholders like your-key-here, xxxxx, <your-api-key>)

    line 145
  8. Distinctive:

    If it has real values:

    line 146
  9. Very specific:

    Create a .env.example that strips the values but keeps the key names:

    line 147
  10. Very specific:

    Delete the original .env file with real values

    line 152
  11. Very specific:

    If .env.example already exists and the .env only has real values, just delete .env

    line 153
  12. Distinctive:

    If any files were flagged as potentially containing credentials but NOT auto-redacted (because they're code files, not env files), call …

    line 193
  13. Distinctive:

    Never delete files outside the target directory

    line 199
  14. Very specific:

    If the project has a .gitignore, use it as a hint — anything in .gitignore is often a good candidate for removal

    line 201

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Sanitize Project for Public Sharing

Clean up a project directory so it's safe and lightweight to share publicly. The goal: remove everything that's either private (credentials, keys) or regenerable (build output, dependencies).


STEP 1: Parse the Path

Get the target directory from the user's argument. If no path was given, ask for one.

Check that the directory exists before proceeding.


STEP 2: Copy First, Sanitize the Copy

Always sanitize a copy — never the original. The original stays untouched with its original name.

# The copy gets a "- Community" suffix (or similar distribution label)
COPY_DIR="${ORIGINAL_DIR} - Community"
cp -r "$ORIGINAL_DIR" "$COPY_DIR"

All subsequent steps operate on $COPY_DIR, not the original. Tell the user:

  • Original preserved at: <original path>
  • Sanitizing copy at: <copy path>

STEP 3: Measure Before Size

du -sh "$COPY_DIR"

Save this number — you'll report it at the end.


STEP 4: Remove Build Artifacts & Caches

These are all regenerated automatically when the project is run — safe to delete:

# Remove each if it exists
node_modules/
.next/
.nuxt/
dist/
build/
.turbo/
.cache/
.parcel-cache/
__pycache__/
*.pyc
.pytest_cache/
.venv/
venv/
tsconfig.tsbuildinfo
*.tsbuildinfo
.DS_Store (all, recursively)

Use find + rm -rf to handle nested occurrences (e.g. nested node_modules).

Example:

find "/path/to/project" -name "node_modules" -type d -prune -exec rm -rf {} +
find "/path/to/project" -name ".DS_Store" -delete

STEP 5: Remove Lock Files

These are regenerated by npm install / pip install etc:

  • package-lock.json
  • yarn.lock
  • pnpm-lock.yaml
  • Pipfile.lock
  • poetry.lock
  • Gemfile.lock

STEP 6: Remove Version Control

.git/
.svn/
.hg/

The recipient doesn't need the git history.


STEP 7: Scan for Secrets & Credentials

Search ALL remaining files for patterns that suggest credentials. This is the most important step — missing a real API key in a public share is a serious problem.

Patterns to search for (case-insensitive):

  • API_KEY, APIKEY
  • SECRET, SECRET_KEY
  • TOKEN, ACCESS_TOKEN, AUTH_TOKEN
  • PASSWORD, PASSWD, PWD
  • PRIVATE_KEY, PRIVATE KEY
  • -----BEGIN RSA PRIVATE KEY-----
  • sk- (OpenAI key prefix)
  • sk-ant- (Anthropic key prefix)
  • Bearer
  • Common cloud prefixes: AKIA (AWS), AIza (Google)

Use grep to scan:

grep -rli "API_KEY\|SECRET\|TOKEN\|PASSWORD\|sk-\|AKIA\|AIza\|Bearer" /path/to/project \
  --include="*.env*" --include="*.json" --include="*.ts" --include="*.js" \
  --include="*.py" --include="*.yaml" --include="*.yml" --include="*.toml" \
  --include="*.sh" --include="*.conf" --include="*.config"

For each file flagged:

  • Read the file
  • Determine if the match is an actual credential value (e.g. OPENAI_API_KEY=sk-abc123) or just a variable reference/placeholder (e.g. OPENAI_API_KEY=your-key-here)
  • Only warn/redact actual values, not placeholders

STEP 8: Handle .env Files

For any .env, .env.local, .env.production, .env.development files found:

  1. Read the file
  2. Check if it contains real values (not just placeholders like your-key-here, xxxxx, <your-api-key>)
  3. If it has real values:
    • Create a .env.example that strips the values but keeps the key names:
      OPENAI_API_KEY=
      NEXT_PUBLIC_SUPABASE_URL=
    • Delete the original .env file with real values
  4. If .env.example already exists and the .env only has real values, just delete .env
  5. Tell the user which files were redacted

STEP 9: Measure After Size

du -sh "$COPY_DIR"

STEP 10: Report

Print a clean summary:

✅ Sanitization complete

📁 Original (untouched): /path/to/project
📦 Community copy: /path/to/project - Community
📦 Before: 458 MB → After: 268 KB

Removed:
  ✓ node_modules/ (458 MB)
  ✓ .next/ (309 MB)
  ✓ .git/ (648 KB)
  ✓ package-lock.json
  ✓ tsconfig.tsbuildinfo
  ✓ 3x .DS_Store files

Credentials:
  ✓ .env.local → replaced with .env.example (stripped real values)
  ⚠️  Possible credential pattern found in: src/config.ts — review manually

Ready to zip:
  zip -r MyProject.zip /path/to/project

If any files were flagged as potentially containing credentials but NOT auto-redacted (because they're code files, not env files), call them out explicitly so the user can review manually.


Notes

  • Never delete files outside the target directory
  • Be conservative: if unsure whether something is a real credential vs a placeholder, flag it for the user rather than silently redacting
  • If the project has a .gitignore, use it as a hint — anything in .gitignore is often a good candidate for removal

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.