paper2code

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

• Third-party content exposure detected (high risk: 1.00). The skill's acquisition and pipeline docs (SKILL.md and pipeline/01_paper_acquisition.md, plus scripts/fetch_paper.py) explicitly fetch and parse public arXiv and OpenReview pages/PDFs and their HTML/metadata, and then use that extracted, third‑party content (including discovered "official code" links) to drive downstream decisions, so untrusted external content can influence agent behavior.