The current behaviour is to exit with a non-zero status when there are some warnings or errors with the examined WordPress installation, like checksum errors.

This is a matter of semantics, but I believe the command should exit with error only if it failed to execute correctly, i.e. wrong flags were given or permission was denied or something else got wrong and the check did not complete. That it's purpose is to find errors and it did, doesn't mean it encountered an error itself.

I plan to use this package in an automated environment and I need to determine when the command completed successfully and when not. The report itself can be seen from the output, if and only if the command completed successfully.