The reason 3DPolylines [and Splines if they are not planar] can't be Offset is that they do not define a plane, and there is no such thing as a "side" to choose, as opposed to "the other side," for a direction for the positional relationship of the result to the source. Your drawing has a lot of 3DPolylines with only the two end vertices, and those are in fact planar, though there are an infinite number of planes any of them could be considered to lie in, spinning around their one segment as a rotational axis.
Would it work for you to Copy them in opposite directions in the context of the current construction plane, as a kind of "pseudo-offset"? It seems that's the situation with your triples -- the Z coordinates of the ends of the ones on either side are the same as each other and as the one between them [in a couple that I checked, at least]. If that will serve, it's probably possible to do that, by determining the two directions to Copy as perpendicular to the XY-only angle of the 3DPolyline. But for those with more than one segment, and some kind of bend(s), what should that determination be based on? The XY-only angle of the longest segment, or something?
Kent Cooper, AIA