The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external resources and dependencies for tool setup:
References an installation guide at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
Suggests installing the belt-sh/cli skill module via the npx skills add command.- [PROMPT_INJECTION]: The skill configuration allows for the processing of untrusted user data, creating a surface for indirect prompt injection:
Ingestion points: User-provided text prompts and media URLs (e.g., image_url, audio_url, video_url) are passed as arguments to the belt CLI tool in SKILL.md.
Boundary markers: Although the input is formatted as JSON, there are no specific agent instructions to ignore malicious directives that might be contained within those strings.
Capability inventory: The skill is authorized to use the belt tool via Bash commands to perform video generation and editing tasks.
Sanitization: The instructions do not define any sanitization, validation, or escaping logic for the content provided by the user.

ai-video-generation