hf-model-inference

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

• Third-party content exposure detected (high risk: 0.90). This skill instructs downloading and loading models from the public HuggingFace model hub (see the "Model Download" steps and pipeline("...", model="model-name", cache_dir=...)), which are user-contributed, untrusted third-party artifacts that the service will load and interpret at runtime.