ai-video-generation
Pass
Audited by Gen Agent Trust Hub on Apr 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs users to external installation instructions for the
infshCLI tool and suggests installing related skills usingnpx, which involves downloading code from theinference-shGitHub organization. - [COMMAND_EXECUTION]: Core functionality is dependent on executing shell commands through the
infshCLI for authentication (infsh login) and video processing (infsh app run). - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it interpolates user-controlled prompts and external URLs directly into command arguments.
- Ingestion points: User prompts and media URLs (image_url, audio_url, video_url) defined in
SKILL.mdexamples. - Boundary markers: No specific delimiters or instructions are used to prevent the agent from following directions potentially embedded in user-supplied content.
- Capability inventory: The
infshtool facilitates network-based video generation and media management. - Sanitization: There is no evidence of input validation or content filtering before data is passed to the CLI.
Audit Metadata