Oh, I'd also test the BMS to make sure it's not the reason the cells died. If it's a passive balancing BMS with a row of resistors that drain the cells to balance them, it's possible that the balancer for that channel is stuck on. If it is, it will then drain the new cells in that group dead over time.
"Easy" test is use a good cell (from whatever source, a new one, or out of some other thing like a tool pack or laptop pack or vape cell, etc) to replace the dead group just for this test. Just measure the voltage of the cell before install, then monitor it over a few hours or days, and if it drops over time, the balancer it stuck on and draining it, so you would also need a new BMS.
Another possible test: with the balance wire disconnected from the dead group, use your multimeter set to the 200ohm or 2kohm range, or continuity / diode test, connect the red + lead to the balance wire, and the black - / ground lead to the side of that channel's balance resistor that is closest to the balance connector. If there is an open circuit (by whatever means your multimeter uses to show this, usually OL or a blank display), then swap the leads and see if it still shows open. If it does, the balancer is probably not stuck on. If it shows a reading in either direction, the balancer is probably stuck on, especially if the reading is very low and about the same in both directions.
Example below of such balancers; you're actually testing the little transistors (Q1-Q12 in this image) to see if they're shorted from collector to emitter.
If you have a good cell that's charged up to at least say, half full, and it's connected to that dead-group's channel, a stuck on balancer will have a voltage across it's resistor (but won't if there's no charged cell attached, and a balancer that's actually off will not have a voltage across the resistor either).
