Kiran-mondal/VisionLangToolkit ? reverse-engineered prompt

Build me a working app called VisionLangToolkit that lets someone pick an image in a Unity interface, send it to a Python backend, and show back the image details in a simple friendly panel.

I want it to feel like a small visual analysis tool, where the user can upload or preview an image and see things like size, color mode, and any other useful attributes the backend can extract. If there is already a place for it, add support for Lua scripts so I can define custom rules for how the image info gets interpreted or transformed.

Please make sure the Unity side, Python side, and any helper scripts all work together cleanly, and include a straightforward way to run the whole project from the repo. If you need to check current docs for Unity or MoonSharp setup, go ahead and look it up online.