You can see how this creates a gray area, since there's no easy way to know what kind of seedling a particular seed will produce, without doing a DNA test. Mangoes and citrus produce seeds with both zygotic and nucellar embryos, and there's no reliable way to distinguish between them.
I am not seeing the grey area so much.
If you grow polyembryonic Mangoes for rootstock ( also polyembryonic Citrus ), it is reasonably easy to tell which are clones for rootstocks.
More so it is fairly easy to tell which are not a clone as they stand out from the rest.
If it really comes down to it, yes you are visually assessing the clones and making assumptions, but this is routinely done in grafting nurseries.
Some pretty close off types certainly get through and some clones are also size graded out.
The issue of propagating a protected fruit tree will probably only arise when it actually fruits.
Ornamental protected plants ie selected foliage types, dwarf hedge varieties etc etc will be breaching rules at any stage. They are protected for visual characters like foliage size colour etc.
Until the protected fruit tree actually fruits it could be difficult to distinguish. The original developer could probably do so based on Mango leaf shape and colour of growth tips etc without a DNA test.
If there is no reliable way to tell which polyembryonic seedling of variety X is a clone, then the reverse is true, there is no reliable way to tell which is not.
Although incorrect, it could be assumed that all polyembryonic seedlings are clones until proved otherwise.
The best strategy is probably quietly grow the seedlings until they fruit. By that time the developer may have move on to newer types also.