video-understand

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

• Third-party content exposure detected (low risk: 0.10). The workflow watches and transcribes videos from arbitrary URLs or local files supplied by the user, ingesting outsider-authored audio and subtitles into the LLM context.

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

• Potentially malicious external URL detected (medium risk: 0.30). The script uses an explicit API endpoint URL to communicate with ElevenLabs during runtime for transcript generation, and ElevenLabs is a third-party service provider without first-party provenance in this context, placing the operational dependency in the medium tier.