voice-ai-development
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides educational code snippets for integrating with well-known voice AI providers. All external domain references target official API endpoints of established services.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a technical surface for processing untrusted audio data from users which is subsequently transcribed for LLM processing. * Ingestion points: Real-time audio streams are ingested via websockets (OpenAI) and transcription callback events (Deepgram) as documented in SKILL.md. * Boundary markers: The code examples do not include explicit system instructions or delimiters to mitigate potential injection within the transcribed text. * Capability inventory: The skill is restricted to network interactions with authorized AI provider APIs. * Sanitization: The instructional patterns do not implement text validation or filtering for the transcribed content.
Audit Metadata