skills/1nfsh/skills/agent-ui/Gen Agent Trust Hub

agent-ui

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Downloads component manifests and SDKs from the official inference.sh domain using verified tools. \n
  • Evidence: npx shadcn@latest add https://ui.inference.sh/r/agent.json.\n
  • Evidence: npm install @inferencesh/sdk.\n
  • Evidence: npx skills add inference-sh/skills@chat-ui.\n- [PROMPT_INJECTION]: Detection of an indirect prompt injection surface due to the nature of agentic chat interfaces. \n
  • Ingestion points: The Agent component ingests and renders untrusted agent responses and tool results streamed through the specified proxy route. \n
  • Boundary markers: Explicit delimiters are not provided in the component setup example. \n
  • Capability inventory: The component supports executing client-side tools (e.g., scan_ui, fill_field) and provides DOM manipulation hooks via formRef. \n
  • Sanitization: Sanitization and security logic are expected to be handled by the @inferencesh/sdk and the associated proxy middleware.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 08:54 AM
Security Audit — agent-trust-hub — agent-ui