xie-shengzi-perspective

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly requires the agent to "先搜索再开口" for fact-finding questions (Step 2: 谢胜子式研究) and lists public third‑party sources (小宇宙播客, B站, 小红书, 即刻/Threads, media sites) as inputs, so the agent is instructed to fetch and interpret untrusted user-generated/open web content that will directly influence its decisions and responses.