gws-sheets

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

• Third-party content exposure detected (medium risk: 0.65). 该技能在运行时会把用户提供的 materialText/materialFile/materialUrl（尤其是 materialUrl 指向的 Sheets 内容或上传文件的文本/OCR）作为可读“表格数据/评论摘要/问题描述”拼入 LLM 上下文；这些内容通常属于操作用户未必选择引入的外部来源，因此存在间接提示注入风险。

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The provided runner sends runtime requests to the external endpoint https://ai-skills.ai/api/execute (default AISKILLS_BASE_URL), which executes the skill remotely and therefore the remote content/response directly controls prompts and execution.