Francis Bond sent me this link for the software used to validate WN-LMF data: https://github.com/globalwordnet/OMW
It would make sense to use this (and other separate processes perhaps) as a part of the release process, e.g. as formal integration tests for the WN-LMF data and for the other output datasets and possibly the DanNet graph itself.