ui-design

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted design briefs and visual media from users. To mitigate potential injection risks, it mandates a journey-first cognitive walkthrough protocol and Gate Reports, which force the AI to validate all inputs against objective design standards and principles rather than following instructions embedded within the data.
  • [EXTERNAL_DOWNLOADS]: The skill includes procedures for downloading brand identity assets. These instructions are coupled with mandatory security checks, including file signature verification and Node.js-based sanitization of SVG files to detect and reject embedded scripts or malicious event attributes.
  • [COMMAND_EXECUTION]: The skill utilizes internal Python and Node.js scripts for search functionality and random style selection. It also leverages standard platform utilities for code graph analysis and HTML export, following established conventions for AI agent skills within its target environment.
  • [SAFE]: The overall architecture is design-focused and secure. It avoids unsafe dynamic execution patterns and explicitly prohibits downloading non-image files, maintaining a high standard of workspace integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 01:34 AM