This PR allows us to run the test suite on a MacOS runner to verify that we get clean test runs. I think we probably want to setup an automation that would kick this off on main periodically and spawn conversations to fix breakages. The build isn't clean on MacOS yet, but hopefully just a few more PRs will get us there... github.com/OpenHands/software-agent-sdk/pull/6039

But for what we do need that? I think runnunig on ubuntu is enough

IMHO any OS that OpenHands runs on is also a legitimate choice for a development environment. Contributors should find clean test suite on any development environment they'd choose to use within the set that OpenHands supports.