Are you dead set against using heating cells and plating? I'm curious as to why. Here is a design that has a cold start, but is utterly stable 10 single cell U, 3.6 efficiency with 1.4 cells per U. It will take auto replace machines (redpower retriever, transposer is what I use) for the isotopes & re-enriched. It heats after it turns on, stabilizes while running at right around 19k heat. 19.6k is the radiation threshold, so it is utterly safe. After it finishes it's fuel it starts to cool again. It doesn't need to be cold to start it again, it should be impossible to explode with infinite runs. If you restart it before it cools off, you will get better returns on breeding. If it cools off all the way, you get 1.4 per U, so still at replacement level. 2 stacks with stack size of 15 heating cells are the minimum to get replacement from that config, 19 gets 2 more cells bred without any additional consequences, and if you are ok with radiation, 23 is the maximum and gets you 2 more cells. Personally, I'm not a fan of radition, so I'd go with stack size 19 heating cells.
I'm not at all dead set against hot reactors; I just don't have any experience with them, and little to no success with my few experiments in the reactor planner. That's a very nice reactor you got there, with more efficiency and output than mine. Of course, it's running hot enough that parking it in the basement is probably a bad idea; a fire-proof reactor hall at some distance seems adviceable. The cell replacement system isn't very hard to build: I used a filter-retriever setup a while back to do the cell swaps on a breeder.
The cold breeder has a few advantages though: less architectural requirements, newbie friendly, and runs just fine without extra mods to do cell swaps. A hot hybrid also needs to run its full cycle; any interruptions will cool it down and ruin the breeding. A hybrid such as this hot design will need 4 MFSUs to receive all that power, unless you actually spend it all during the cycle - and I don't. A cold breeder on the other hand can pulse to top up an MFSU without any breeding loss, giving excellent continuous power with no effort. (though I recommend a second MFSU inline between reactor and main mfsu to catch the overflow energy that resuts from the reactor shutting down too slowly)
Seeing how you managed to get a 10-core reactor up though, I'm now curious and may have to start experimenting with hotter reactors.