In the current implementation, proof states are not updated after proving new conjectures! Therefore, subsequent proof attempts are unable to take advantage of such proved conjectures. This needs to be fixed.