The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the infsh command-line interface to execute image processing workflows. It uses infsh login for user authentication and infsh app run to trigger remote models like Real-ESRGAN and Topaz. These operations are standard for the tool's intended use.
[EXTERNAL_DOWNLOADS]: The skill recommends using npx skills add to fetch additional functionality from the inference-sh/skills repository. Because this repository belongs to the official vendor (tul-sh), these references are documented as safe vendor operations.
[PROMPT_INJECTION]: The skill accepts user-provided image URLs for processing, which establishes an attack surface for indirect prompt injection (e.g., instructions hidden within the image metadata or content).
Ingestion points: The image_url field within the JSON input for infsh app run commands across the SKILL.md file.
Boundary markers: Absent.
Capability inventory: The skill can execute remote image processing applications via the infsh tool.
Sanitization: Absent; the skill passes the URLs directly to the external inference.sh platform for processing.

image-upscaling