Score could be kept with citations. You’d be required to list the work you built on, as we do today, and the authors would receive credit. No citation would be worth more than another. If you published something useful for a particular field or made a major discovery that opened a new field, then your citation count would reflect it.
Wouldn’t you be able to game that by having 2 entities spamming citations for each other?
If grants are tied to the “score” there is an incentive to abuse the system.