That sounds like what I remember from discussions in the last few weeks. They updated the help pages on this, but it almost seems like as many questions have been raised as have been answered.
Ideally, you really want all parties on GEDmatch and WikiTree, of course, but it seems like that's unusual to be able to get. Yet it's the ONLY way to properly prove what you're claiming so anybody else can look it up and verify it.
It sounds like they're not really requiring that. On the one hand, that's dealing with practical reality. On the other hand, it puts the confirmation on the same basis as a published genealogy that reports on what family members told the author when he contacted them - you're asked to believe they're honest and competent.
It sounds to me like they just want the initials, due to the GDPR. It seems to me like the person should also be identified with their relationship to their closest living relation in the line, but I'm not sure if that's even OK. For example, while you're saying "I" (initials only) is a 2C1R, you would think it would helpful to it to say that they were also a child of "H" (if "H" is deceased). But does that count as "additional identifying information" in violation of the privacy rules? I'm not sure.
Since this is a 3C or closer, it is NOT what they call a triangulation. In THAT case, and they're less rigorous about it. I don't think you even need to get them over on GEDmatch so you can identify a segment. The easiest criticism of the help page criteria is that it sounds like a GEICO ad ("save up to 15% or more" - which literally could mean ANY amount onf savings, or none at all). It says something like that it has to be "approximately" at AncestryDNA's "3rd Cousin" level (which is from 90cM to 200cM, BTW). Now there's only just certain levels, and they're all just cM intervals, so what they really need to do is say that the match has to be at least "X.cM". What they have in this regard is literally meaningless - 3Cs can go all the way down to 6cM on Ancestry DNA.
Anyway, since it's not triangulation, I guess they have a bit more faith in the paper trail, and just assume that if you have a paper trail to A+B than chances are that you really ARE biologically related to both A & B. Or maybe they figure that the higher cM count that you're seeing is from multiple segments, and thereby very likely to have come from both gt-gt grandparents. But, again, they don't really even give a required cM level, or number of segments, last I saw.
It'd be nice if they shared WHY the criteria is what it is. As it is, it could be a few random cronies of unknown qualifications, getting together in a cozy little corner of cyberspace and making decisions based on either their own personal experiences, or using a Ouija board - for all I know!