resume-match

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

• Third-party content exposure detected (medium risk: 0.65). 该技能在运行时会“抓取 URL 的页面内容/对图片做 OCR 识别/读取外部简历文件内容”，这些输入可能包含他人（outsider）撰写的自由文本，从而被拼入 LLM 上下文（例如 JD URL 的网页正文、简历中他人附件/下载内容的文字）。

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches and parses a user-supplied JD URL at runtime ("如果是 URL，抓取页面内容"), so arbitrary remote page content (i.e., the user-provided JD URL / any HTTP(S) URL) would be injected into and directly control the agent's prompt/context.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)