ml-paper-writing

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs the agent to perform WebSearch and arXiv searches and to use third‑party APIs/services (e.g., arXiv search URLs, Google/Google Scholar via WebSearch, Semantic Scholar/CrossRef/arXiv APIs and the Exa MCP endpoint https://mcp.exa.ai/mcp) to fetch PDFs, metadata, and verify BibTeX, so the agent will ingest and interpret untrusted public, user‑submitted content as part of its workflow.