web-article-translator

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

• Third-party content exposure detected (low risk: 0.10). 该技能在工作流“获取网页内容”步骤中使用 mcp__web_reader__webReader 读取用户提供的 URL 的页面内容并将其以 Markdown 形式交给 LLM 翻译，因此存在外部作者文本被模型直接摄入的间接提示注入风险。

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly uses a web-reading tool to fetch arbitrary webpages at runtime (e.g., https://example.com/article), and the fetched page content is injected into the model for translation, so remote content can directly control the agent's context and enable prompt-injection or execution of instructions.