higgsfield-food-beverage
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely descriptive and instructional, focused on crafting high-quality prompts for food and beverage video generation. No malicious logic or command execution patterns were found.
- [SAFE]: The skill implements a mandatory human-in-the-loop requirement, stating that generated prompts and credit costs must be presented to the user for explicit confirmation before any tool execution occurs.
- [SAFE]: No indicators of prompt injection, data exfiltration, or obfuscation were detected in the instructions or reference materials.
- [SAFE]: Tool access is limited to the Higgsfield service-specific tools defined in the configuration.
Audit Metadata