azure-ai-voicelive-ts
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Downloads official Azure SDK packages from npm, including @azure/ai-voicelive and @azure/identity, which are well-known and trusted services.\n- [EXTERNAL_DOWNLOADS]: References official documentation and source code hosted on the Microsoft Azure GitHub repository.\n- [PROMPT_INJECTION]: The skill inherently possesses a surface for indirect prompt injection because it processes external audio and text input from users or function outputs.\n
- Ingestion points: Defined in SKILL.md through event handlers like onResponseAudioDelta, onResponseTextDelta, and onInputAudioTranscriptionCompleted.\n
- Boundary markers: None explicitly defined in the provided implementation examples.\n
- Capability inventory: The skill enables bidirectional WebSocket streaming of audio and text data.\n
- Sanitization: The provided code follows standard event handling without custom input validation or sanitization logic.\n- [SAFE]: Recommends the use of DefaultAzureCredential for authentication, which is a security best practice for managing identities in Azure applications.
Audit Metadata