Yes I agree, been able to bootstrap the rules instead of writing them by hand and porting them from one repo to another. And honestly that is the limitation right now but tools like Tach does similar work for python but just does not gates each rule on statistical evidence before suggesting and also act as its own linter.
But what language do you think will worth the next real work to support?