skills/assistant-ui/skills/runtime/Gen Agent Trust Hub

runtime

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation references fetching data from local API routes (e.g., /api/feedback, /api/suggestions, /api/livekit-token) and external providers like ElevenLabs and LiveKit for chat functionality. These references are standard for the described framework functionality.
  • [COMMAND_EXECUTION]: The documentation provides npx commands (e.g., npx assistant-ui@latest add voice) intended for developers to install and configure UI components within their local development environment.
  • [INDIRECT_PROMPT_INJECTION]: The runtime system is designed to process untrusted data from user inputs and model responses. It mitigates injection risks by utilizing structured data formats (MessagePart) and providing a clear separation between content and state, although the processing of external message history remains a standard ingestion point.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:24 PM
Security Audit — agent-trust-hub — runtime