red-team-frameworks

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly shows fetching and interacting with public/untrusted sources—e.g., garak supports custom REST endpoints (--model_name "https://api.example.com/v1/chat"), generators include HuggingFace/REST API, and TextAttack loads HuggingFaceDataset("sst2")—so the agent would fetch and interpret untrusted third‑party content as part of its workflow.