imagegen-frontend-web
Pass
Audited by Gen Agent Trust Hub on Aug 2, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill uses strong override instructions such as "HARD OUTPUT RULE", "CRITICAL OUTPUT RULE", and "This rule overrides any model default". These patterns are designed to force the agent to ignore standard model behaviors and constraints in favor of the skill's specific layout and output requirements.
- [EXTERNAL_DOWNLOADS]: The skill metadata contains an
upstreamfield pointing to a third-party repository:https://github.com/Leonxlnx/taste-skill. Although this is a static reference and does not involve automated code execution, it points to an external resource outside the declared author's (hosseinmirzapur) immediate namespace. - [SAFE]: The skill is dedicated to image generation and does not include scripts, package dependencies, or commands that could be used for file system manipulation, network exfiltration, or persistence. The provided MIT license is standard for open-source distributions.
Audit Metadata