skill-management

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill searches external repositories on GitHub and marketplace platforms using gh search and fetch, ingesting third-party repository/skill descriptions and contents at runtime.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (medium risk: 0.30). The script scripts/search.ts fetches data from https://skills.sh at runtime to retrieve and parse marketplace listings.