voice-ai-engine-development

Warn

Audited by Snyk on Feb 28, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.80). The skill's WebSocket conversation flow (see SKILL.md and examples/complete_voice_engine.py
  • the /conversation endpoint) accepts client audio, transcribes it, and feeds those untrusted user-generated transcriptions into the LLM agent, which directly drives synthesis and next actions, so arbitrary third-party/user content can influence agent behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 28, 2026, 11:49 AM