Interesting thought exercise.
I think you would need to flesh out assumption gamma quite a lot more. One of the classic fundamental assumptions is the universal uniformity of physical laws.
Since the deviation from flat hasn't really been a part of the model, that would mean that the vast emptiness of interstellar distances that represent nearly all of the observable Universe would operate under apparently different rules because they are empty and nearly flat ... but because they are dark the alternative rules are not observable.
It has been years, but I've read a bit about the topic. Measurements at a visible distance through red shift and similar tend to suggest that at least within the measurable universe the physical laws are invariant. When it comes to visible features, what we see in a lab matches what we see in the solar system, which matches what we see in our galaxy, which matches what we see in distant galaxies and what we see in stellar lenses. The inaccuracies are within measurement error. We don't discover distant galaxies with fundamentally different levels of gravity where orbits no longer fit the classical rules. We don't discover stellar lenses where gravity is somehow different in a region, providing a different lens property than what we expect. ... Except...
... Except those are for things that we can see. Those are in non-flat areas. The flat areas we cannot see so they could operate differently.
Introducing a scaling factor based on matter or spacetime in dark/flat areas may work out or not, I don't know that much about the physics involved. But the thought that it could happen is interesting.