I cannot speak for the particular cell pack in the ELT but having dismantled and rebuilt quite a number used outside aviation it is standard to include a fuse or a polyfuse (effectively a self resetting fuse) in a pack. These often turn up as links between the actual cells of multi cell batteries but on single cell designs they are at the cell ends of the leads or on an attached protection circuit board.
Whoever did that design would have been expected to test their design (i.e. apply both increasing and random levels of excess discharge current and finally an immediate short circuit across the battery terminals or wires) to ensure the battery shut down and stayed shut down safely in each case.
From the reports the two leads were shorted together which should have resulted in a very short pulse of current then a drop to a hugely limited current for the duration of the fault thanks to the action of the polyfuse. If the polyfuse failed to work or failed to reduce the fault current enough for safety the cells would then be able to potentially overheat over time running the risk of Cell thermal runaway. If the short circuit was intermittent this might have been the subject of many repeated cycles of short, shutdown and recovery and perhaps this is where the issue arose
Once the cells reach a critical internal temperature (c 135c IIRC) they proceed to break down and self ignite and no fuse is going to help them after that point