thought-mining

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

• Third-party content exposure detected (high risk: 1.00). Stage "第三阶段：观点验证" (stages/03-validation.md) explicitly instructs the agent to perform "联网搜索" and record reference links, so the agent will fetch and read open/public web content (third‑party sources) as part of its workflow, exposing it to untrusted user-generated content.