Some unit tests should get written, if the publishing of releases becomes automated as suggested in #8. The GitHub action for publishing a new version should only run, if the tests pass. The less code is covered, the higher the chances are for broken versions to be released.