linuxmirrors-awesome

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

• Third-party content exposure detected (medium risk: 0.65). SKILL.md 的工作流要求在 Step 3/4 “按权威来源读取官网、文档、仓库、更新日志和安全公告”并进行交叉验证，这会在运行时拉取/读取外部公开网页与 GitHub 等内容（outsider-authored free text），从而进入 LLM 上下文。

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires at-runtime verification of dynamic facts by fetching the project website and GitHub (e.g., https://linuxmirrors.cn/, https://linuxmirrors.cn/use/, https://linuxmirrors.cn/other/, and https://github.com/SuperManito/LinuxMirrors), so these external pages are used during runtime to supply content that will be ingested and control the agent's outputs.