I found a company to precisely fix this problem. Every dataset, every analysis, every computation fully transparent, reproducible, accessible. You can even embed it in your blog. For instance check out https://igor.sbgenomics.com/lab/public/pipelines/ if someone uses one of those, and share their task, there is full transparency + instant replication.