talking-head-b-roll

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

• Third-party content exposure detected (medium risk: 0.30). The workflow ingests outsider-authored audio/video content through Whisper or ElevenLabs during transcription (transcribe.mjs), which processes untrusted speech containing free text.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill fetches a Whisper model from Hugging Face via curl during setup instructions, representing an unofficial remote dependency.