huggingface-papers

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

• Third-party content exposure detected (medium risk: 0.30). Required runtime workflow ingests outsider-authored free text because user-supplied paper IDs/URLs can lead the agent to fetch paper content/abstract as markdown via https://huggingface.co/papers/{PAPER_ID}.md (and/or the HTML fallback on arXiv), and that content is then available to the LLM for summarization/analysis.