alterlab-hypogenic

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs cloning public datasets from GitHub (e.g., https://github.com/ChicagoHAI/...) and placing external research PDFs in literature/YOUR_TASK_NAME/raw/ to be processed via GROBID and pdf_preprocess.py so those untrusted, third-party materials are ingested and used by the HypoRefine/Union pipelines to generate and refine hypotheses that directly influence model prompts and downstream inference.