Due to the issue described in #7393, we end up re-requesting these values again in pull requests since we generate filters that do not contain these values. An easy work around is to maintain a separate no-contact-info table of crds hashes, very similar to the purge table the CRDS db maintains.