I'm going to try to rewrite our linters to emit minimal fixes and try that
out. I do think that will make the conflicts nearly nonexistent.

I think adding the "requirements of a custom linter" and possibly also
adding verification of original text matching the byte range for fix
application should still be addressed, though.

It feels like the basis of the problem is that linting was written well without autofix in mind and then the feature was added later. It wasn't designed in from the beginning and that results in limitations where compromises were/are necessary. Unfortunately, this means users sometimes end up with unexpected or unintuitive behaviors. I think we're all just trying to find ways to mitigate them.

Here's an example:

This appears to still be open.