The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references installation instructions and related tools from the official inference-sh GitHub repository and website. These references are part of the standard setup and usage for the integrated service.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as image URLs and text prompts, which is typical for image processing tasks.
Ingestion points: Arguments passed to the belt CLI (image_url, prompt) as shown in SKILL.md code examples.
Boundary markers: No specific delimiters or safety wrappers for the injected content are defined in the command-line examples.
Capability inventory: The belt CLI performs network operations to interact with the remote inference API to process images.
Sanitization: Explicit sanitization logic is not defined within the skill instructions; validation and sanitization are expected to be handled by the underlying CLI tool and the remote API service.

background-removal