Ability to use problem matcher results to determine success/failure of Action steps

Currently the success/failure result of a GitHub Actions step when using shell scripts (run key) is determined by the exit code of the last command.
I’d like to be able to ignore the exit code and let the result only depend on the problem matcher results, i.e., let it pass if there are no problems found and fail if some exist.
I need this because of python/mypy#10013.

Would there be any options or workarounds to achieve this?
Otherwise, please consider it as a feature request.