This is so unique! So I have a few notes that I would like to cover based off of experience.
Firstly I have some concerns about the gear. From a completely unbacked opinion, it is my inexpert opinion that the ring gear will need to be manufactured out of aluminum.
What we found with our old modules is that not only will the teeth on the ring gear sheer, the bearing track will deform, allowing the gear to flex out of the way and be damaged. Even with aluminum, this is a risk at extremely small levels which cause excess wear, or worst-case, lead to skipping.
Fresh gear on left, used on right
Taking a look at our first module’s gears after testing, you can notice that the internal teeth aren’t there anymore. 
Although this is an extreme case (we used homemade bearings, what else was gonna happen), many of our new modules equiped with West Coast x-contact bearings have a tendency to shower aluminum dust if gears are misaligned even a smidgen.
Other remarks can be made about the rigidity of the module. For the top and bottom plates on your module, I would recommend a more solid method of connecting the two than standoffs. We found on our modules that the two plates were subject to immense torsional force that would twist them against each other, causing the gears of the differential to misalign.
This can be solved by replacing the top and bottom plates with aluminum and adding some form of 3d printed standoff with matching pocketed registers on the top and bottom plates.
Observe the black 3d printed spacer in the center of the module
Another remark I have are the geared ratio. I have calculated the turning ratio as 10:1, (MK4 is 21.4), which is a bit low, but still impressive considering the setup. Having the drive ratio 100/10 * 36 / 80 = 36 / 8 = 4.5 is so commendable, however, the ratio also seems low.
4.5:1 with 8 simulated modules on ILITE Simulator returns a 26ft sprint
As the field dimensions are 25x54, the adjusted speed of 16fps will be hard to reach on a regulation FRC field.
Differential swerve has two engineering challenges that define a decent module, gear ratios and power draw.
Looking a the simulation results, we can see that the gear ratios are not perfect, however what is more troubling is that the motors are browning-out. This was an issue that Mars Wars encountered on their differential swerve. Our team chose to counteract this obstacle by designing a 3-wheel chassis. (if we are breaking the taboo of diffy swerve, why not also do triangle?)
4.5:1 with 8 simulated modules on ILITE Simulator returns a 18ft sprint
With an increased adjusted speed of 16.4, we can see that in addition to decreasing current-limiting. We also reduce our sprint time, creating a more viable drivetrain.
Another thing to note is the modularity of our module design. This is EXTREMELY SUBJECTIVE, but our team went with modules that are easily dismountable from the chassis, as well as an easy to remove wheel, which allows us to swap both the module and the wheel in pit without any hassle.
Our modules are built modularity so we can easily remove partial parts of our module without removing the mounted module
Differential stack completely removed from module
We also have a 1/4-20 bolt that holds 3/8 hex shaft with our wheel
We created a quick-disconnect wire-harness for the module for hassle-less module swappage
Obviously this is extremely subjective because it truly depends on the level of your paranoia.
Overall, this module design has so much potential. It is actually a perfect candidate for the next big drivetrain revolution and a step to bringing differential swerve to commercial availability. The ingenuity of placing a motor in the center of the module makes the module look as if its coaxial is sort of full-circle.
