research-assistant

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs agents to search and ingest content from public third‑party sources (e.g., SKILL.md and references/research-assistant.md: "문헌 조사 계획: 국내(RISS, DBpia) 및 해외(Google Scholar, PubMed)" and "DB 탐색: Google Scholar, arXiv, PubMed, SSRN 등 검색"), then reads and synthesizes those materials into analyses and decisions, which could allow indirect prompt injection from untrusted/web-sourced content.