I agree: if a taxonomic name change hasn’t undergone peer review it shouldn’t be adopted here, and passing peer review doesn’t immediately justify accepting a change.
Further to the discussion of different kinds of peer review, it’s important to understand what peer review actually evaluates. In general, peer review assess the scientific rigour of a paper. That might include the appropriateness of the analysis, the sufficiency of the data, and how the results are interpreted.
It doesn’t necessarily apply to the taxonomic conclusions. The science may be rock-solid, but it won’t necessarily follow that there is a single correct taxonomic outcome. Different scientists may take the same data and interpret it as supporting two species, two subspecies, two varieties, or two forms that don’t warrant formal names. The peer review process may filter out more extreme claims, but species vs subspecies sorts of things are unlikely to have a firm yes/no answer.
There’s no official rule or governing body to adjudicate such issues. If you check a few online references you’ll see that they don’t all agree in all cases, and none of them are actually ‘official’ (for plants at least, I think some of the animal groups may actually have official naming authorities?).
In practice, such disputes are resolved informally when the majority of scientists adopt one or the other of the competing options. That process used to take years, as the new results filtered slowly through journals into the hands of specialists, and eventually into field workers. Nowadays it appears to take days, given everything is available more or less instantly online.
All of which to say there’s a very strong case to be made that iNat ought to take a very conservative approach when changing taxonomy. In most cases waiting a year or more after a publication before adopting a change would be reasonable in my view.