gpt-taste
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The instructions establish a professional design persona and enforce strict layout and motion rules without attempting to bypass agent safety protocols or extract system prompts.
- [COMMAND_EXECUTION]: The skill includes instructions to manage a local design wiki by creating and appending to markdown files within the llmwiki/ directory. These file operations are restricted to the workspace and follow a predictable documentation pattern for tracking design iterations.
- [DYNAMIC_EXECUTION]: The requirement to simulate Python script results is a prompt engineering method to generate layout variance and does not involve the actual invocation of external code or shell environments for malicious purposes.
- [EXTERNAL_DOWNLOADS]: The skill references standard image placeholder services (picsum.photos) and well-known UI libraries, which are routine in frontend development and do not represent a security risk.
Audit Metadata