web-fetch-cce

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

• Third-party content exposure detected (high risk: 0.85). 该技能的运行时“Web Fetch”会对运行时获得的任意 URL 抓取网页内容（HTML/inner_text/Markdown），属于公共网页等外部来源经由 WebFetch/Crawl4AI/Playwright 解析后进入 LLM 上下文的间接提示注入路径。

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches arbitrary external webpages at runtime and injects their content into Claude prompts (e.g., fetching WeChat articles such as https://mp.weixin.qq.com/s/...), so remote page content can directly control the agent's instructions/context.