OK, now I understand what you are talking about.
Our primitive architectural descriptions of buildings and parts when applied to individual biomolecules and their folding may not be so helpful or even accurate sometimes, but to talk about an RNA structure that has parts that function like “beams” (very stiff) and parts that function like “hinges” (highly flexible) and other types of parts and all highly dependent on this complex interdependent interaction, I am still forced to use a general word “scaffold” for the stiff-well-defined-structural parts whether you approve of it in your domain or not.
Additionally, even when talking about the structure you cite above (STE5), when we break it down into how each of the pieces operates, we are still forced to talk on the nanometer scale about units of structure that are like beams or blocks and flexible parts that work like hinges. One also sees that whereas you have a complex architecture in this protein (with 700+ aa), it is not composed of multibranch loops, it is the standard combinations of alpha helices and beta strands separated by several large regions of flexible, functional semi-non-structured (when not binding a target) features.
My view of a real scaffolding protein is something like NS1 that wraps up influenza A viral RNA segments. That is really functioning purely as a “scaffold” in the most literal sense of our human-derived word. For STE5, I would still have chosen a word like “processing protein”. Scaffold wholistically implies that the STE5 is simply a dumb passive filing cabinet rather than a highly complex interacting piece of machinery.