Give me some examples of science where we make decisions and keep the raw data a secret. I can’t think of any except ones involving national security.

Yes, the best way to prove reproducibility is to reproduce it, but that can’t always be done and at any rate is no excuse for keeping raw data a secret.