hugging-face-vision-trainer

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and inspects public, user-uploaded Hugging Face datasets (e.g., load_dataset("username/dataset-name") in references and scripts and the dataset_inspector.py which calls the Datasets Server API at https://datasets-server.huggingface.co/rows?dataset=...), and those dataset fields (including prompt/annotation columns) are parsed and used to drive preprocessing and training decisions, so untrusted third-party content can directly influence tool use and workflow.