image-to-video
Warn
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs users to install external tools and dependencies using unpinned community packages such as
npx skills add belt-sh/cliandnpx skills add inference-sh/skills@ai-video-generation. These installations from non-whitelisted sources introduce risks associated with unverified supply chain components. - [COMMAND_EXECUTION]: The skill relies on the
beltCLI tool within a Bash environment to perform multi-stage workflows, including remote model inference, media upscaling, and audio merging. - [INDIRECT_PROMPT_INJECTION]: \n- Ingestion points: The skill processes user-supplied text prompts and local image file paths through the
belt app runcommands. \n- Boundary markers: No explicit delimiters or boundary markers are used in the prompt templates to distinguish between instructions and potentially malicious data embedded in user inputs. \n- Capability inventory: The skill is grantedBash(belt *)permissions, which allows for network-based content generation and local media processing. \n- Sanitization: There is no evidence of sanitization or validation for user-provided prompt strings or file paths before they are interpolated into the command-line arguments.
Audit Metadata