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

elements

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs users to fetch component source code and styling metadata from the official r.assistant-ui.com registry.
  • [COMMAND_EXECUTION]: Provides standard developer instructions for using the npx assistant-ui and npx shadcn command-line tools to initialize and manage project components.
  • [INDIRECT_PROMPT_INJECTION]: The UI components documented in this skill, such as Thread, MarkdownText, and ToolCall, are designed to ingest and render untrusted data from LLM message streams and external tool outputs.
  • Ingestion points: Data enters the system via message part streams processed by Thread.aui, MarkdownText, and specialized reasoning/source renderers.
  • Boundary markers: The skill utilizes the AssistantRuntimeProvider and underlying primitives to scope and delimit content within the chat context.
  • Capability inventory: The components support rendering rich text, markdown, Mermaid diagrams, and interactive tool approval cards.
  • Sanitization: Content sanitization is managed by the underlying @assistant-ui/react primitives during the rendering lifecycle to prevent malicious payload execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:50 PM
Security Audit — agent-trust-hub — elements