community-publish

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

• Third-party content exposure detected (medium risk: 0.30). The skill reads open-sourced code catalogs via list_open_source() and downloads external project files via fork(), which ingest outsider-authored free text from the community GitHub repository.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill fetches file listings over HTTPS from the public GitHub API repository Starchild-ai-agent/community-projects at runtime, which constitutes an unofficial external operational dependency.