My current repository houses only atomic findings and I find they have remained pretty timeless thus far. I’m struggling to scale it because of a number of things, but one of my rules is to only put live product in it. This is partially because prototype behaviour isn’t as real as that you see on released products, so it’s a little less reliable. But also, product teams tend to run their own prototype testing and I’d probably amend their approaches if I had the time to do so. I’m a believer in bad research being better than no research but pragmatically, you need to be cautious about the ‘findings’ becoming folklore in the company.