Yep, the spindle should also be extended whenever the beast is NOT heating the rads or on overrun -see below.
The low resistance measurement is predictably inconclusive.
The problem is clearly when the flow switch is ON not when it's off, so the high measurement is not an issue.
You could pull both connectors off the flow switch and if they're 1/4 inch spades, use a paperclip to connect them together.
However, if the voltage across the wires stays low when the ch starts getting hot the flowswitch doesn't appear to be at fault. This is another benefit of measuring the voltage - to measure the resistance you'd have to turn it off. If the voltage is correct when the behaviour isn't, that isn't a snapshot.
The boiler would be expected to circulate after through the CH after a HW demand, to dissipate excess heat, but not for long. You're seeing that.
It's possible that the sensors are telling the boiler it has overheated, but I'd have thought it would turn the burner off!
The remaining possibles include the other circuit board(s), which is/are involved in turning CH on (overriding hw) as mentioned above. Broken tracks on potentiometers can do odd things.