I’m not saying that I “know better” than Gould or MOLA, I certainly don’t. I’m saying that the kinds of cues I’m using in the example here are fundamentally different in ways that I clearly laid out than the kinds of cues imagined by the MOLA guide. They are on a different staff. If they were on the same staff as the players part, I would agree that they should be transposed as though they were the flute part.
I agree that nearly all cues should be transposed for the part they appear in, and that should absolutely be the default. However, I think this is a reasonable example of an instance where they might not make sense transposed. These cues are only for coordination in passages where rhythm is either free or complex; they appear on a separate staff; and one of the parts is for voice. Another concern is that this is a large transposition (up a fourth), and so it creates a lot more ledger lines than it would to leave the part at sounding pitch. The bottom line here is that this whole part is much more like a score that has some special affordances to be played by the flute player than most parts.
Last, I’m not trying to convince anyone here that this should be the way they show cues, or even that what I want to do is ok, just that this should be something that is allowed by the software, which should enable users to do whatever it is they are trying to do, not what it thinks they are trying to do.