remotion-best-practices

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

• Third-party content exposure detected (high risk: 0.80). The skill's calculate-metadata workflow explicitly shows fetching arbitrary URLs (fetch(props.dataUrl) in rules/calculate-metadata.md) and using the returned JSON to set composition props and duration, meaning untrusted third-party content is ingested and can directly change tool behavior.

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

• Potentially malicious external URL detected (high risk: 0.70). The transcribe-captions guide uses @remotion/install-whisper-cpp (see https://www.remotion.dev/docs/install-whisper-cpp) and its installWhisperCpp/downloadWhisperModel calls in a runtime Node script to fetch and install whisper.cpp and model files (remote code/artifacts that get executed), which is a runtime fetch that results in executing downloaded native code.