red-blue-energy-training-ppt
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is text transformation for image generation, with no executable code, network operations, or sensitive file access patterns detected.
- [PROMPT_INJECTION]: Instructions focus on maintaining visual consistency and content extraction. There are no patterns attempting to bypass safety filters, extract system prompts, or override platform constraints.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to generate image prompts. It mitigates risk by instructing the agent to 'refine and structuralize' the content into a specific template rather than directly interpolating raw user text, which reduces the surface area for prompt-based manipulation of the image generator.
- [COMMAND_EXECUTION]: All tool usage is limited to the platform-integrated
image_genutility for its intended purpose.
Audit Metadata