The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external resources, including installation instructions at raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md and related skills from the inference-sh and belt-sh GitHub organizations.
[COMMAND_EXECUTION]: The skill executes the belt CLI tool via Bash to interact with the inference.sh API. Access is restricted to this specific tool through the allowed-tools frontmatter configuration.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it processes user-provided prompts and external URLs (image/audio) which are then passed to the video generation models.
Ingestion points: User input via the prompt, image, and audio fields in the JSON input for the belt app run command.
Boundary markers: The skill uses structured JSON to pass parameters but lacks specific instructions to the agent to ignore potentially malicious content within the user-provided prompt.
Capability inventory: The skill is capable of executing the belt CLI tool to communicate with remote inference servers.
Sanitization: No explicit sanitization or input validation logic is present in the markdown instructions; the skill relies on the downstream CLI and service-side safety protocols.

p-video