I really like the idea of the project but most of the github copilot models do not have vision support. I am using playwrighter with opencode and it works great, just the screenshots break sessions since the model will no longer respond with a screenshot in context