To me the #1 thing that fixes the social issues described in this post (changing code someone checking in code a few hours after they pushed it) is a code review process. The most valuable thing about always doing pull requests and code review is sharing information and having discussion about changes, not necessarily the immediate improvement of the code. A team will gain more knowledge about what everyone is doing by reviewing each others pull requests than they do from a daily standup by a pretty large margin in my opinion.

In this particular example, reviewing a PR would have meant that a discussion about whether or not that abstraction was a good change before the code was ever merged.