doc-coauthoring

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

• Third-party content exposure detected (medium risk: 0.30). Source: SKILL.md runtime workflow; it ingests outsider-provided free text when “如果用户提供了素材文件：直接读取…提取关键信息” and when users reply with topic/reader/style material during Context Gathering and refinement, which is user-authored content fed to the LLM without selecting a specific pre-authored item first.