selfagency/opilot ? reverse-engineered prompt

Build me a VS Code extension that brings Ollama into GitHub Copilot Chat so I can use local or remote models right inside the editor.

I want an Ollama sidebar in the activity bar where I can see my models, pull new ones, start and stop them, delete them, and open a simple settings view for things like temperature and token limits. It should also show a status bar indicator so I can tell when Ollama is running and whether it’s online.

Inside Copilot Chat, I want to be able to pick an Ollama model from the model selector or type @ollama to chat with it. Please make it support streaming responses, tool calling, vision input when the model allows it, and inline code completions from a chosen model. Local models should stay on my machine, and if I point it at a remote Ollama server it should work with that too. If you need current docs, look them up online.