media-crawler

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). Suspicious — the GitHub URL points to a third‑party/unvetted repository that instructs you to clone and run install scripts (bash/python) which can execute arbitrary code on your machine; the 127.0.0.1:8080 URL itself is just a local WebUI endpoint and not an external download, but it would be served by whatever code you run from that repo.

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md and README explicitly run MediaCrawler (e.g., "uv run main.py --platform xhs/dy --type search") to scrape public social platforms like 小红书/抖音/微博/贴吧/知乎, which ingests user-generated, untrusted web content into the skill's runtime (saved under data/{平台}) and could influence downstream decisions or actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The setup script performs runtime fetch-and-execute of external code—specifically it runs "curl -LsSf https://astral.sh/uv/install.sh | sh" to install uv and clones/updates the repository "https://github.com/NanmiCoder/MediaCrawler.git" which is then executed (uv run main.py), so external content is fetched at runtime and used to execute code.