muapi-ui-design
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user descriptions and includes them in prompts sent to the image generation backend without adequate sanitization or boundary enforcement.
- Ingestion points:
scripts/generate-mockup.shvia the--descargument. - Boundary markers: Uses
[UX_BRIEF]and[EXECUTE]tags, but lacks explicit instructions to the AI to ignore instructions within the user-provided context. - Capability inventory: The script invokes a shell subprocess to run
core/media/generate-image.sh. - Sanitization: None; user input is interpolated directly into the prompt string.
Audit Metadata