storyboard-creation
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill demonstrates usage of the
beltCLI tool for remote image generation and image stitching. It explicitly requests access to this tool via theallowed-toolsfrontmatter field. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-defined descriptions to generate image prompts, which is a standard feature for this type of agent skill.
- Ingestion points: Scene and shot descriptions within the command examples in
SKILL.md. - Boundary markers: None provided in the command templates.
- Capability inventory: Limited to
beltCLI tool invocations for API-based image processing. - Sanitization: No specific sanitization or escaping of prompt input is mentioned.
Audit Metadata