He couldn’t. All he could do would be to dictate the code of conduct for physicists working at the MPI. And if you don’t like his code of conduct, you can go to a different institute, or take it up with the community or the board of directors for the institute. Your implied scenario of “what if a nazi wrote a code of conduct” doesn’t scare me because the only thing worse than a nazi is a nazi in hiding. And if the code of conduct contains no nazi agenda, well then he can be held accountable if he does something against the code of conduct (like, for example, being a nazi)

But back to your point. Your point was made using two examples. Bad examples, in my opinion, since the first doesn't apply (consumption of work vs collaboration) and the second one also doesn't apply (building on others work vs collaboration).

Your further point about watching what people do vs what they say they do is a reasonable point, but doesn't have much to do with the value of a code of conduct. I'm pretty much in agreement, that's the best way to proceed when collaborating with a single individual - but communities are more than the sum of their parts. If there are three people collaborating on a project, and one swears a lot in their commits, and one prefers a lack of swearing, and one doesn't care either way - how do you resolve that conflict? Nothing you've said offers a useful tool to navigate that situation.