ui-design

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFE
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill is designed to produce and verify runnable UI artifacts (HTML, CSS, and JavaScript). It leverages browser tools such as browser_navigate, browser_vision, and browser_console to load, visually inspect, and interact with these agent-generated files. This is a standard and necessary component of its UI design and prototyping functionality.
  • [COMMAND_EXECUTION]: Included in the skill are utility scripts (evals/run-static-checks.sh) used for validating the skill's file integrity and structure. These scripts utilize common system utilities like jq, grep, and python3 to perform static analysis and ensure the skill meets its internal structural requirements.
  • [EXTERNAL_DOWNLOADS]: The skill metadata references established and trusted design and accessibility resources from organizations such as the W3C, Apple, Google, and the Nielsen Norman Group. These references are used as authoritative guides for visual craft and usability standards.
  • [SAFE]: The instructions incorporate proactive security measures, particularly regarding the postMessage architecture used for host integration. It recommends origin pinning to mitigate cross-origin risks and explicitly mandates that no sensitive user data or identity information be stored in the skill's operational memory.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 02:21 AM
Security Audit — agent-trust-hub — ui-design