youtube-digest

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs and shows examples of passing API keys as command-line arguments and writing them into plaintext config files (e.g., --api-key ak_你的密钥 and echoing JSON with the api_key), which requires handling and potentially outputting secret values verbatim.

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

• Third-party content exposure detected (high risk: 0.75). scripts/extract.py 的运行时路径会把调用外部 API 返回的 YouTube transcript/metadata（尤其字幕文本，属于非操作者来源的内容）渲染并直接写入 LLM 上下文：通过 render_lines() 形成可读纯文本 lines，并随后 print/build_markdown 输出这些文本；这些文本可被宿主代理当作模型输入（潜在间接提示注入）。