You are making an interesting point. However, I'm not sure NOMA does indeed make these assumptions. It restricts the area in which science and religion have the right to work. Science, however, has already restricted itself by its principle. NOMA implies that there is an area for religion to work in - science does not, because it does not concern itself with matters outside its area. What is the problem?
The restriction science places upon itself are based upon philosophical ideas at to how we can ever access 'truth' (if indeed we can access it at all). Science restricts itself to that which is testable (or falsifiable) because it recognises that those are the only claims that we can ever have any insight into the truth of. Take, for example, and unfalsifiable claim - we'll use in this case that an invisible, immaterial and undetectable elephant sneezes on people and makes them itch. What does a universe containing this unfalsifiable entity look like? Compare it to a universe that does
not contain this unfalsifiable entity. What differences are there between the two worlds?
The answer, strangely enough (or not if you were expecting it), is that there is no difference. When one posits an unfalsifiable entity, one is describing an object that necessarily has no influence on the observable universe (because if it did have an influence we could test for it, and it would no longer be unfalsifiable). One could imagine that the entity influences the
unobservable universe, but then one runs into the same issues of justification as before: How do we know an unobservable universe exists if we can't, by definition, observe it? And, following on from that, what does it even mean to say that such an object or place exists, given that we can never know of its existence, and that it can not influence the observable universe in any way?
Certainly there are issues that science cannot tackle - ethical systems, while they can be
studied by science, can be devised and/or codified only by philosophers. If science were to devise an ethical system, it would run afoul of the rule that one should not obtain a normative statement from a descriptive statement. In much the same way, a philosopher can develop an ethical system that he believes to be perfect, and a scientist could study how well the system actually works in practise. And the boundaries are not clearly defined - there is blurring around the edges, and many people play dual roles as scientists and philosophers (though that does not make the science they do philosophy any more than it makes the philosophy they do science).
But where for religion to fit in? Put simply, it doesn't. Religions claim to provide answers to the questions posed by philosophers and scientists, but they fail even the most basic levels of inquiry. That is not to say that religion has never done any good - it obviously has - but where it has done good it has not done so because any religion has truth to it, but instead because it has provided a united community, and in that there is power; power that can be used for good or for evil...and most commonly for both. But this does not make religion necessary, and it does not lend credence to the truth of any religion.
That is why the NOMA argument fails. The domains of the magesteria of science and philosophy overlap in many areas, thus violating the idea of 'non-overlapping magesteria', and the magesteria of religion necessarily violates both the magesteria of science and philosophy while not itself having any established claim to truth.
