faceforensics_classifier

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs downloading pre-trained models from the public GitHub repo https://github.com/ondyari/FaceForensics, so the skill consumes third-party (user-hosted) model artifacts that are used at runtime to make classification decisions and could therefore materially influence behavior.