wechat-cover

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

• Third-party content exposure detected (medium risk: 0.30). 运行时从红狐接口 https://redfox.hk/story/api/gzh/search/hotArticleNew 取回 articles 中的 imageUrl、以及 title/author/url/summary 等字段后，报告模板会把这些（尤其是 d.keyword 与 p.case.title/author/url/p.prompt）注入到可执行的 HTML/JS 渲染页面并打开浏览器。

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

• Potentially malicious external URL detected (high risk: 0.80). The skill's runtime script calls the external API at https://redfox.hk/story/api/gzh/search/hotArticleNew to fetch article/image data that are injected into the generated report and used to build image-generation prompts, so fetched content directly influences agent prompts at runtime.