remotion-video-toolkit

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

• Third-party content exposure detected (high risk: 0.80). The skill instructs the agent to fetch and ingest arbitrary external URLs (e.g., rules/calculate-metadata.md uses fetch(props.dataUrl); rules/assets.md, rules/images.md, rules/videos.md accept remote Img/Video/Audio URLs; and rules/lottie.md fetches remote Lottie JSON), which are untrusted third‑party resources that the agent is expected to read/interpret and that can modify props/metadata and thereby influence rendering decisions — creating a clear vector for indirect prompt injection.