gpt-taste
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of instructional guidelines focused on frontend design engineering. It does not contain any executable scripts, data exfiltration logic, or sensitive credential exposure.
- [COMMAND_EXECUTION]: The skill instructs the AI to "simulate a Python script execution" within its internal design plan. This is a cognitive prompting technique intended to randomize layout choices and does not involve the execution of actual shell commands or system-level processes.
- [EXTERNAL_DOWNLOADS]: The skill references well-known, legitimate services for UI assets, such as using
picsum.photosfor placeholder images. These are standard tools for frontend development and do not pose a security risk. - [REMOTE_CODE_EXECUTION]: While the skill directs the AI to use libraries like GSAP and Phosphor Icons, it does so in the context of generating code for the user, rather than downloading or executing untrusted code at runtime within the agent's environment.
Audit Metadata