voice-ai-development
Pass
Audited by Gen Agent Trust Hub on Apr 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains boilerplate code for interacting with OpenAI, Vapi, Deepgram, and ElevenLabs. All sensitive values like API keys are represented by safe placeholders (e.g., 'sk-...', '...').
- [SAFE]: Network operations are restricted to establishing connections with the official API endpoints and WebSockets of well-known AI service providers for transcription, synthesis, and model inference.
- [SAFE]: The technical instructions focus on latency optimization and user experience best practices, such as implementing barge-in detection and streaming audio, with no evidence of obfuscation or unauthorized system access.
- [SAFE]: Dependency usage is limited to standard Python packages and official provider SDKs required for the documented functionality.
Audit Metadata