The cube test is an interesting one, but, to pile on another stupid internet trend, that seems more like a "shape-rotator" vs. "word-cel" question. For very simple fold-ups I can do this, but for more complex decompositions (like if you have me an octahedron) I doubt that I could reliably do this.

In terms of the triangle attaching, is it that without drawing you would not be able even to say what the resulting shape is? If I threw in a third triangle you wouldn't be able to tell me what the shape is now? I'm trying to see if I can get a clear discriminator that tells me whether I am different from you in how I am able to visualize or if we're just using different words (or if I, too, am aphantashic and that other people have powers beyond my ken).