wechat-article-review

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

• Third-party content exposure detected (high risk: 0.75). 该技能在运行时会把用户提供的 materialText/materialFile/materialUrl（尤其是 materialUrl 对应的公开文章正文）作为可读文本进入上下文；若这些内容来自非操作用户（如外部链接抓取的文章正文或他人草稿/文档），则属于“OUTSIDER-AUTHORED FREE TEXT”经由上下文合成进入 LLM。

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

• Potentially malicious external URL detected (high risk: 1.00). scripts/run.py makes runtime POST requests to the external execution endpoint at https://ai-skills.ai/api/execute (default AISKILLS_BASE_URL https://ai-skills.ai), meaning the remote service executes the skill and thus controls prompts/responses and is a required runtime dependency.