adversarial-machine-learning

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

• Third-party content exposure detected (low risk: 0.10). The required workflow fetches external content from a GitHub raw README file, which is an outsider-accessible repository containing community-submitted links and text.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs fetching external behavior-selecting content from an unauthenticated third-party GitHub repository (gmh5225/awesome-ai-security), presenting a weak/unknown provenance operational dependency.