The core feature is a one-click way to mute a randomly failing test. Muting is different than skipping a test because a muted test will still run on CI, but it won't fail the entire build if it fails. Keeping a flaky test running in your CI is usually valuable so that you gather more data about it and that your error logs stay relevant.
It only works with pytest for now, but I'll add support for other test frameworks if there is any interest!