It's a good idea: have a fully-open repository where negative results can be documented. With the ability to tag on if a negative result is replicated.

Then gradually re-enculture review boards to expect any given author to have a substantial number of results published to said repository, if they're actively performing research.