voice-ai-development

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides templates and patterns for building voice agents that process untrusted audio and text data from external users, which is a known attack surface for indirect prompt injection.
  • Ingestion points: Data enters the agent's context through the vapi_webhook Flask endpoint and the transcribe_stream function (SKILL.md).
  • Boundary markers: The provided code snippets do not include prompt delimiters or instructions to ignore embedded commands within the processed transcripts (SKILL.md).
  • Capability inventory: The skill demonstrates capabilities for network communication with AI provider APIs (OpenAI, Deepgram, ElevenLabs, Vapi) and includes patterns for tool/function calling (SKILL.md).
  • Sanitization: The implementation patterns do not demonstrate sanitization or validation of the generated transcripts before they are passed to the language model (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 07:20 AM
Security Audit — agent-trust-hub — voice-ai-development