We used to use PC's for production testing and found they would periodically lockup. It was always attributed to "ESD" but never proven.
Anyway I have a suggestion for what I think will be a productive test. While certainly you need some impedance between you contacts and the µP you can't be sure the offending signal is only bbetween the input wire and ground.
My test suggestion: Short the switches to the "closed" position on / near the processor board. Operate it in the field for enough time that a lock would have likely occurred.
- If the processor locks up then you have a common mode problem (i.e. signal coming in on both wires). Note this test does not prove there is no normal mode issue as well.
- If the processor does not lock up then its a good indicator the problem is normal mode and RC filtering is likely to be helpful.
In addition to the ability of the processor to not lockup in "normal" operation, such a setup must be able to withstand many ESD discharges into your wires or the nail heads etc. Think of a 100 pf capacitor charged to 6K or 8K volts being connected into every wire going into and out of your processor box. There are tougher standard out there but it gives you a sense of the capability of the suppressors need to have.
Everyone has their own design thoughts, personally I prefer ESD zener diodes at the entrance of external inputs.