liucixin-perspective

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

• Third-party content exposure detected (high risk: 0.80). The skill's required Agentic Protocol (SKILL.md, "Step 2: 刘慈欣式研究") explicitly instructs the agent to "use tools to verify" scientific claims and the "调研来源" section names public media and web sources (e.g., NYT, Goodreads, news sites), indicating the agent will fetch and interpret open/public third-party content during normal workflow, which could materially influence decisions and actions.