Based on Slarty and Robert Rapplean's answers and comments concerning The Cluster World I'm seeing a far more active world than I originally envisioned, some short notes follow.
Atmospheric retention is unlikely to be an issue over far longer timescales than I had considered and the cluster will be increasingly hot with depth due to frictional heating during the atmosphere's introduction.
There is likely to be a relatively narrow zone of human habitable atmospheric pressure at some depth below the surface of the Cluster.
A percentage of the core of the cluster will be heated by friction between air molecules during the migration of the atmosphere leaving a zone of semi molten hot rocks. Over the two millennia of the original question they will, probably, remain glowing hot and semi-fluid to the surface, that's simply not enough time for shedding that much thermal energy (it is possible that the rate of convection lose is high enough to cool the core significantly but almost certainly not completely).
There will also be increasing atmospheric pressure with depth, this means that the cluster would have needed significantly more gas than I had originally thought to have a surface atmosphere but I'm okay with there being no air on the surface. The added atmosphere would have a volume of roughly 1.23x1021m3 at STP.
The atmosphere in and around the super-heated core zone is likely to be a liquid rather than a gas for some depth due to pressure, assuming there is sufficient water there will be a boundary layer defined by the surface of a boiling ocean.
The structure will be rotating to even out surface-down heat effects but there is no specific rate of rotation set at this time.
The rocks won't vapourise no matter how much thermal energy they absorb. This is an unusual artifact of the way they are pinned in place. I'm not sure if this means they'll stay in place at temperatures that should make them boil away and/or puts a hard limit on their maximum temperature.
Here is a, very simplified, cross section to help understand what I believe the prevailing conditions once the system has reached initial dynamic equilibrium will look like. NB: The rocks are significantly tighter packed than shown in this cross section, 70% of the total volume is rocky material, this is demonstrative of the gradients in effect not the actual structure:
The temperature and pressure gradients are from highest to lowest, white hot in the core out to blackbody solar equilibrium at the outer edge, the liquid (solid?) core ocean with a liquid air layer and then a gas phase atmosphere that fades out to vacuum.
Now I'm interested in the hydrological implications of this situation, in particular will water vapour be able to circulate up out of the high pressure core zone into the habitable pressure zone, or would the pressure in the lower atmosphere put a lid on water convection with the upper levels only seeing dry gas plumes?
I appreciate that the scenario in this question is right on the edge of modern scientific knowledge, thus I have left off the science-based metatag but as in the original question I would like to stick to known science as much as possible. In short while this environment was created by magic its ongoing behaviour is governed by physics.
Thanks again to the workshopping team's help over in the Sandbox your help is absolutely invaluable this close to the edge.