to fill with PU foam before applying the plaster? [...] Or does that result in any disadvantages
As already says, don’t do any foam nonsense there. Between stones goes mortar and nothing else. Additionally, I would cover that with mesh, as with chases.
the masonry joints (which were not closed due to tolerances in structural engineering)
Joints because of tolerances? – no, I am clearly for “zero tolerance” and even for the introduction of corporal punishment. Because I suspect you are the victim of one of those botched planners who have no idea about stone dimensions and construction reference measurements and just know the wall thicknesses. And who then, for example, distribute the excessive 3 cm somewhat evenly across the butt joints in a 2.77 m wide wall section. That gives in the example an average of just over half a cm play per joint, thanks to the interlocking you can’t see straight through, but professional execution looks different: the plan stones are intended for dry butt joints – but not to interpret the butt joints as a crumple zone/play. Obviously, heat and sound rush through there.
The masons are not to blame, stoning the planners is not their responsibility, and special compensation stones for planners ignoring reference measurements do not exist. Therefore, the masons have no choice but to bridge reference measurement misses smaller than thin-format stones with such improvisation.
Leaving air joints would present similar plaster crack risks as with foaming (both are improper fillings for the joints). I see mortar as the only suitable means; the working hours for that of course must be charged to the “planner.” As far as we are talking about joints of one and a half cm, mortar alone is sufficient; for joints of three or four cm I would already apply mesh.
Just for the misuse of the term “construction tolerances” for such purposes, I would let the “planner” build the next house entirely alone. Then he would have learned it. Is the garage already built? – he could practice nicely there.