I love that you ran the numbers. I think you've been very aggressive; I would expect the battery being discussed (RV energy storage, not a traction pack) to operate at well under 1C normally and have much lower heating than your worst-case scenario.
Here are some counterpoints, to further the debate -- which I think is a great one, btw:
1) Aluminum shell cells need to be insulated from each other for electrical reasons, regardless of the thermal situation. So we are having a somewhat academic debate: especially if the manufacturer provides spacers that allow for an air gap, the air gap is "free" modulo volume concerns in the RV. It would be strange to design a different packaging method that removes that air gap. (Now, I think they design in the gap because many of their packs
are used in traction applications, where there will surely be substantially higher heating. But this does not alter the fact that in many cases there will be an established, manufacturer-approved install method that already provides for the gap.)
2) Your calculation finds a worst-case 28degC increase. I would argue that even a 5degC increase distributed unevenly throughout the battery pack is unacceptable. We know that heating is one of the biggest contributors to aging. We know, empirically, that others with LFP packs experience more cell degradation on the interior cells when they are arranged in rows or grids. In an RV, there likely will be times when the enclosure's ambient air temperature reaches or exceeds 40degC. I would not want the middle of my pack to be a concentrator of whatever heat does accumulate in that case, if I could help it, because the degradation scales nonlinearly with temperature. Does a real pack generate even a 5degC delta in the middle under mild loads? I don't know for sure.
3) Your calculation does not account for any ohmic heating at the interconnect. Those can be substantial -- especially once wire, dissimilar metals, mistakes or wear (under-torqued connectors), and especially fuses become involved. All of those potential heating sources can transfer heat right back into the cells if they are in close contact. (My own DC system's total resistance is roughly 25mOhm!)
I have personally measured, with a FLIR camera on a single bare aluminum shell cell in open air at room temperature, at 0.25C for about one hour, a gradient of just under 1degC between the middle of the largest exterior face of a cell and its outer edges. I could easily imagine that becoming five or ten times that when densely packed with little convective opportunity. I haven't done any more analysis, because, well, see #1 above. But I think it is too early in our understanding of how this chemistry behaves in the long tail to simply dismiss outright possible heating issues in a large battery pack operating at lower rates. Yes, the heat here is second-order compared to a traction application... but some of us are trying to baby these cells and stretch well into the many thousands of cycles that it looks like this chemistry can deliver, under ideal conditions. Why risk it by introducing suboptimal conditions?