The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses templates that interpolate user-provided descriptive text into shell commands for image generation.
Ingestion points: User-defined prompt strings, subjects, and styles within the Bash examples in SKILL.md.
Boundary markers: None; user input is placed directly inside JSON objects passed to the belt command.
Capability inventory: Access to the belt command via Bash, as defined in the skill's allowed-tools configuration.
Sanitization: The skill does not provide mechanisms to escape or sanitize user input before it is processed by the shell.
[EXTERNAL_DOWNLOADS]: The skill references external installation resources and provides instructions to add related skills.
Fetches setup documentation from the inference-sh organization's public GitHub repository.
Recommends installing the belt-sh/cli skill and other related skills using the npx skills add command.
[COMMAND_EXECUTION]: The skill provides numerous Bash examples to run AI models via a specific CLI utility.
The execution scope is correctly restricted using the allowed-tools frontmatter to only permit the belt command.

logo-design-guide