elon-musk-perspective

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

• Third-party content exposure detected (high risk: 0.90). Yes — the SKILL.md Agentic Protocol (Step 2: "⚠️ 必须使用工具（WebSearch等）获取真实信息，不可跳过。") explicitly requires the agent to fetch and interpret open web content and then use those facts to drive decisions, which exposes it to untrusted public third‑party pages that could contain indirect prompt-injection instructions.