A Claude Code skill that checks whether AI agent skills are safe before you install or use them.
claude install fysoul17/pyx-scan/pyx-scan owner/name
Example:
/pyx-scan anthropic/web-search
Queries the PYX Scanner API and returns a safety verdict:
- SAFE — Scanned and verified safe
- OUTDATED — Previously safe, but updated since last scan
- CAUTION — Potential risks, requires your confirmation
- FAILED — Flagged as dangerous, do not use
- UNSCANNED — Not yet scanned
MIT