product-ad-shot-design-by-1d090cw
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes prompt excerpts and case summaries from
references/cases.md. It utilizes a structured 'reference contract' (Preserve, Replace, Avoid) and an 'evidence gate' to ensure the agent grounds its output in the provided data while maintaining clear boundaries between reference material and user-supplied brand context. - [EXTERNAL_DOWNLOADS]: The skill includes references to public media and source links on
goodcase.aiandx.comfor visual reference. The workflow explicitly instructs the agent to use available inspection tools (image, video, or browser tools) to verify these sources, following standard creative research practices. - [SAFE]: The instructions include explicit safety and attribution rules, such as prohibiting the creation of unauthorized accounts, preventing the impersonation of creators, and requiring user approval for any external changes or uploads.
Audit Metadata