line-voice-agent

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The documentation provides an installation command for the Cartesia CLI from the vendor's official infrastructure (https://cartesia.sh).\n- [REMOTE_CODE_EXECUTION]: The curl-to-shell pattern is used exclusively for the installation of official vendor tooling.\n- [DATA_EXPOSURE]: Sensitive API keys are managed securely using environment variables and the vendor's encrypted secret management system.\n- [SAFE]: The skill provides detailed patterns for mitigating indirect prompt injection risks associated with processing user speech input:\n
  • Ingestion points: Transcribed speech from UserTextSent events.\n
  • Boundary markers: Instructions provided via system prompts in LlmConfig.\n
  • Capability inventory: Network search, telephony actions, and custom tool execution.\n
  • Sanitization: Examples include regex validation and multi-agent content filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 04:44 PM