Many of us have heard or, or experienced a Class X cap failing (RIFA), but how many have had a double pole rocker switch fail shorted, or even arc between the poles when used at it's nominal rating? It has to be a spectacularly rare event, if it has happened at all.
If someone smashes the switch then they could directly expose the live connection, a far more serious problem than possible short to the other pole, so that's another irrelevancy. As previously mentioned, this isn't even a consumer device.
Those switches regularly fail, and I have had them blow up as well when moisture provides a bridge, or the little springs inside decide to break and the floaty bits short out the 2 sides. The plastic also degrades with time, and as the inner side often is made from recycled or mixed content ( the bits that were either rejects repelletised, or the recycled plastic of mixed colour from some random seller) as it is not visible. Stove repairers are very well aware of how poor these switches are, they are a common failure on many things, either popping the top off exposing the live inner parts, or simply shorting, or even just falling apart with time. Hard to tell the quality unless you have a well known manufacturers unit, and that could be replaced by anybody with a non correct part for either cost cutting, or as a service part, without anybody being the wiser, as the cutout and pin outs are a common industrial size.
I agree on the use of a 22.5mm industrial control panel button and using some industrial contact assemblies, as those are both robust, double acting fail safe contacts and each module is isolated from each other to withstand 2.5kV for 30 seconds insulation test, just place 4 blocks so the wear is kept the same, even if one block is never going to be used, and so that the mains section is separated physically from the 12V side by a barrier strip.