arxiv-monitor

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly fetches and ingests papers from the public ArXiv API ("https://export.arxiv.org/api/query" in Step 2) and writes their summaries into a digest that the morning briefing loop reads and acts on, so untrusted third-party paper content can influence agent decisions.