termination-review

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

• Third-party content exposure detected (high risk: 0.90). Yes — the workflow's Step 3 ("管辖地特定要求") and the "研究连接器预检" explicitly instruct using legal research connectors (e.g., yuan dian MCP or other configured research tools) and optionally "搜索网络" / produce [联网检索——需复核] results, meaning the agent will fetch and read public web/third‑party sources whose content is used to determine compensation, timing, escalation and next actions.