video-ad-specs
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs users to download installation instructions from an external GitHub repository (
inference-sh/skills) and install a CLI tool from a non-trusted NPM organization (belt-sh/cli). - [COMMAND_EXECUTION]: The skill provides numerous shell command examples using the
beltCLI to perform tasks such as logging into a remote service and executing AI model inference for video, audio, and caption generation. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of video ads by interpolating user-provided prompts into model execution commands. This represents a potential surface for indirect injection if the input content is not properly sanitized before being passed to the underlying AI models.
Audit Metadata