the spark came from right on the edge of the tab, there was a black mark there,
I think it must have shorted via the plastic/foil , it was the aluminium tab.
below shows pic of where the sparks occurred, black circle shows where the bolt head was located ( it overlapped onto the plastic), red dot shows where the sparks were. It seems only possible that the short was via the plastic ( I doubt there was anything metal on the table at the time it occurred) note the hole I've punched in the cell pictured is a different cell not the one that shorted
I worked out a way to join the tabs together that seems to work quite well and very simple/quick process.
The cells I have , have very short tabs

What I've done is fold one tab, slip the other tab into it and squash them together, that alone is not sufficient,
so I then used a tiny punch ( screwdriver in my case) to hold the copper/aluminium in place.
View attachment 1
It makes quite a secure joint for low amp applications ( in my case about 20amps max). Once the tabs are joined the cells need to be taped together as any movement between the cells will weaken the joint.
When I do the rest of the joints I'll use a bit of vaseline inside the joint, and some roofing silicone to keep moisture out.
I think it will last ok for my use, as long as the cells cant move against each other it should be ok.
I do need to pack between the tabs when I'm doing the punching which is a bit of work.
I've used some kapton tap to stop any shorting occuring and some water proof tape over the kapton.