image-to-video
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill references tools and repositories owned by the vendor (inference-sh), such as the 'belt' CLI and related skills hosted on GitHub.
- [EXTERNAL_DOWNLOADS]: It provides instructions for downloading configuration and installing CLI tools from the vendor's official GitHub organization and package registry.
- [COMMAND_EXECUTION]: The skill guides the agent on using the 'belt' CLI for authenticated sessions and media processing tasks, which are standard operations for this service.
- [INDIRECT_PROMPT_INJECTION]: The skill handles user-supplied text and image paths. While this represents a potential surface for indirect injection, the skill defines a rigid command structure for the agent to follow.
Audit Metadata