This always had the smell of a low tension problem.
But what is quite intriguing is how, if the only thing that was ever done to the instrument was a bit of a clean-up at the points, the insulation arrangement of the engine-stop brush could have altered.
With a meter, the brush should show no continuity to the alloy with the cover off, measuring from the carbon or from the stud/tag for the stop wire.
With the cover on, you'd expect a very low resistance reading from the stop wire to the bike, representing the 1/2 ohm resistance of the low tension winding (from cb centre screw head to main magneto earth), plus any small resistance across the brush and spring.
If the bike ran perfectly before, and if the problem is the simple fault described, the bike should run again with the brush removed or freshly insulated, using any suitable washer in fibre or plastic.