deepgram-sdk-patterns

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests arbitrary public URLs (see "Step 3: Audio Intelligence Pipeline" analyzeConversation(audioUrl: string) which calls transcribeUrl with { url: audioUrl }, and the Python transcribe_url method), transcribes and summarizes/derives intents from that untrusted third-party content, and those derived outputs can materially influence downstream behavior.