mux-docs

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

• Third-party content exposure detected (high risk: 0.85). This skill’s runtime workflow explicitly fetches public Mux documentation markdown (e.g., https://www.mux.com/llms.txt and per-page .../page.md) and then feeds that fetched free-text into the agent for answering, which is outsider-authored content via runtime HTTP retrieval.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs the agent at runtime to fetch external docs/markdown and answer from them (e.g., https://www.mux.com/llms.txt and https://www.mux.com/docs/guides/start-live-streaming.md and https://raw.githubusercontent.com/muxinc/elements/refs/heads/main/packages/mux-player/REFERENCE.md), so remote content is fetched during runtime and directly controls the agent's responses.