The Agent Skills Directory

[SAFE]: The skill uses established procedures and bundled Python scripts to manage server discovery. It explicitly advises against using potentially dangerous shell commands like find, ls, or which, instead favoring controlled API interactions.
[INDIRECT_PROMPT_INJECTION]: The skill provides an interface for ingesting data from external ComfyUI servers.
Ingestion points: The skill reads server info, system stats, node schemas, and model inventories (SKILL.md, Procedure steps 1, 5, 8, and 9).
Boundary markers: Not explicitly defined in the markdown instructions.
Capability inventory: Command execution is limited to specific bundled scripts (python scripts/comfyui_control.py and python scripts/workflow_catalog.py).
Sanitization: Processing logic is contained within the referenced scripts, minimizing prompt-level injection risks.

comfyui-discover