❓ Frequently Asked Questions (V3)

🔒 Security & Trust

Q: What do the Risk Labels mean?

⚪ Safe (White/Blue): Read-only, planning, or benign skills. Safe to run anywhere.
🔴 Risk (Red): Skills that modify files, delete resources, or perform security scans. Use with caution.
🟣 Official (Purple): Maintained by trusted vendors (Anthropic, DeepMind, etc.).

Q: Can these skills hack my computer?

No. Skills are just text files (markdown instructions). However, they instruct the AI to run commands. If a skill says "delete all files", your AI might try to do it. Always review the code before creating a skill, and check the Risk label.

📦 Installation & Usage

Q: Why shouldn't I just copy all files?

You can! But 250+ files might clutter your context window. We recommend using Starter Packs (docs/BUNDLES.md) to install only what you need for your specific role.

Q: Does this work with Windows?

Yes, but some "Official" skills use symlinks which Windows handles poorly by default. Run git with: git clone -c core.symlinks=true ... Or enable "Developer Mode" in Windows Settings.

🛠️ Contribution

Q: My PR failed "Quality Bar" check. Why?

V3 introduces automated quality control. Your skill might be missing:

A valid description in the metadata.
Usage examples. Or it might be triggering a security filter. Run python3 scripts/validate_skills.py locally to see the error.

Q: Can I update an "Official" skill?

No. Official skills (in skills/official/) are mirrored from vendors. Open an issue instead, and we will forward it to the maintainers.

1.7 KiB Raw Blame History