zhang-yiming-perspective

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

• Third-party content exposure detected (high risk: 0.90). This skill's required Agentic workflow (Step 2 in SKILL.md: "⚠️ 必须使用工具（WebSearch等）获取真实信息" and concrete examples like "先WebSearch小红书...") instructs the agent to fetch and interpret public web content (news, interviews, social media, and arbitrary URLs listed in references), which are untrusted third‑party sources that can materially influence the agent's decisions and tool use.