tools-ui

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions suggest using the shadcn CLI to download UI component configurations from https://ui.inference.sh/r/tools.json.
  • [COMMAND_EXECUTION]: The documentation provides shell commands for users to install the belt-sh/cli and other related skills such as inference-sh/skills@agent-ui.
  • [INDIRECT_PROMPT_INJECTION]: The ToolResult and Agent components ingest data from external tool responses to render them in the UI. 1. Ingestion points: The result prop of the ToolResult component and the tools configuration in the Agent component. 2. Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the provided UI snippets. 3. Capability inventory: The components are designed for UI rendering; no direct file system or network capabilities are shown in the component code snippets themselves. 4. Sanitization: There is no evidence of sanitization or filtering of the tool output before it is passed to the display components.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 04:09 PM
Security Audit — agent-trust-hub — tools-ui