We do have blind spots: we only run the tests on high density cells for example (though this is more because of compute resource limitations- would be alleviated by self-hosted runners which we're looking into), and we don't have full coverage of all the tools: dr-cu and cu-gr come to mind. We're gradually improving though, and PRs are appreciated. :) Other things we'd like to tackle include native aarch64 Docker support for all the M1 Macs out there.