ml-model-integration

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly queries and ingests user-generated content from the public HuggingFace Hub (via HfApi.list_models and by loading/running models with pipeline), and those third-party model metadata and model outputs are read and used to shortlist, evaluate, and decide deployments—so untrusted external content can materially influence agent actions.