ai-video-generation
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user prompts and external URLs (images, audio) to generate video content.
- Ingestion points: User-provided prompt strings and remote asset URLs (
image_url,audio_url) are passed to theruncomfytool inSKILL.md. - Boundary markers: Inputs are encapsulated in a JSON structure (
--input '{...}') before being passed to the CLI, which reduces shell injection risks. - Capability inventory: The skill uses the
runcomfyCLI tool to perform network operations and write generated files to the local file system (--output-dir). - Sanitization: The skill relies on the
runcomfyCLI and backend service for content safety and input validation; no specific local sanitization of prompt text is performed by the skill itself beyond JSON structure. - [EXTERNAL_DOWNLOADS]: The skill facilitates the installation and use of the
@runcomfy/cliNode.js package and communicates with officialruncomfy.comandruncomfy.netdomains. These resources are provided by the authorruncomfy-comand represent standard vendor functionality for the service described.
Audit Metadata