verification-rules

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly requires generating verification documents after "웹 조사 완료" (web research completed) and "NTIS 검색 완료" (NTIS search completed) for multiple chapters, which indicates the agent consumes public web/third‑party search results that could influence subsequent actions and thus exposes it to untrusted third‑party content.