I think the issue is that it's a 7S pack because of the ++ -- (10p) connection in the middle. He would have been trying to charge at 8s voltage, and could have easily overcharged, especially with the mis-matched parallels, and no BMS ?. All this of course is assuming that the black dots mean either positive or negative, I don't know that for sure and can only assume.He probably would have caught one mistake at the end of the string but two mistakes made the ends come out right. My gut feeling says it just wouldn't work, not start a fire. We should work out the math on how much juice could flowed through the tiniest bottleneck. If even possible, how much would it take and what would happen if you forced flow backward through cells? What reverse voltage (charging) did those cells see?
I think the mistake at the beginning would have only severely limited capacity and usability to the lowest cell. The voltage should have been the same and overcharging "shouldn't" have occured (of course there's no BMS either). Will did a test of an active balancer where he put different capacity cells in series; it worked, and I don't think it was a fire hazard. However the capacity is limited to the smallest cell in the series, which in his case would have been the single cell at the beginning and ends of the pack.
Again, that is assuming that he didn't add another bus bar before he hooked the pack up at the end, he very well could have put one on when he hooked up the charger / loads.