deepgram

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

• Third-party content exposure detected (high risk: 0.90). The skill's "Pre-Recorded Transcription" section explicitly shows transcribe_url({"url": "https://example.com/meeting.mp3"}) which fetches arbitrary public audio, and the Streaming Transcription example/comments indicate those transcripts are sent to an LLM for response generation—meaning untrusted third-party audio (and its transcription) can be ingested and influence agent actions.