The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external resources for functionality, including installation instructions for the required CLI tool hosted on a remote repository (https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md) and the installation of a dependent skill (belt-sh/cli).
[COMMAND_EXECUTION]: The skill heavily utilizes the belt CLI tool via Bash to interact with remote AI models, perform authentication, and execute image processing tasks as part of the storyboarding workflow.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided text to generate visual content, creating a surface where instructions embedded in the data could influence the agent's behavior.
Ingestion points: User-provided visual descriptions and scene details used for image generation prompts in SKILL.md.
Boundary markers: Absent; there are no clear delimiters or instructions for the agent to ignore potentially malicious content within the user-provided descriptions.
Capability inventory: The skill has access to the belt CLI tool via Bash(belt *), allowing it to run arbitrary AI models and stitch images (SKILL.md).
Sanitization: Absent; user input is interpolated directly into command strings without escaping or validation.

storyboard-creation