ai-video-generation
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill leverages the 'belt' CLI to perform video generation tasks. It utilizes commands such as 'belt login' for authentication and 'belt app run' to execute specific AI models with user-provided parameters like prompts and URLs.
- [EXTERNAL_DOWNLOADS]: The instructions reference installation scripts and documentation hosted on the official GitHub repository for 'inference-sh'. These resources are required for the setup and operation of the 'belt' tool.
- [PROMPT_INJECTION]: The skill processes user-supplied text and image URLs to generate video content. This data is passed as input to the 'belt' command. While this represents a surface for indirect injection via processed data, the scope is restricted to the specific API parameters of the inference tool.
- [SAFE]: All identified external domains (inference.sh) and GitHub repositories (inference-sh/skills) are associated with the primary service provider. No obfuscation, persistence mechanisms, or unauthorized credential access patterns were found.
Audit Metadata