I'm unsure what you mean exactly.

I'll assume you mean the trust from people who are not directly involved with protocol changes (endusers, ..) in those people who propose changes to the protocol/system.

"Trust verification" has always been a messy process, I'd clame, and it will always be. Every individual will form their opinion whether they can trust the system as a whole or a certain element of the system or another individual person (developer, ..) on the basis of a diverse mix of information sources, past experience and intuition. It's just how society works.

The best path IMO would be:

- implement a few (maybe a handful) more simple (!) changes

- then completely stop modifying the system because the greater the complexity grows, the likelier it is that unforeseen instabilities/breakages can result

Complexity is a killer. For a system that's interlinked with human society to remain dynamically stable, it has to be kept below some threshold complexity, because otherwise the share of people who're able to understand it's system dynamics will become too small. If only a small fraction of people understand a system they cannot behave rationally in its context which leads to instability.