wewrite-learn

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md 运行路径中，用户可触发“学习这篇文章/导入范文 + URL”为 wewrite fetch-article <url>，随后内容作为 /tmp/article.md 被 wewrite exemplar /tmp/article.md 读入并注入写作模块的初稿 prompt；该路径会把第三方网页抓取的自由文本（URL 站点正文）带入 LLM 上下文，属于 public web content fetched at runtime。

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

• Potentially malicious external URL detected (high risk: 0.90). The skill accepts and fetches arbitrary external content via the user-supplied URL ("") at runtime (e.g., wewrite fetch-article <url> ... and wewrite learn-theme <url> ...) and then injects that fetched content into exemplars/prompts, so external content can directly control the agent's prompts.