I wouldn't expect bonding done to NEC standards to prevent the high voltage arcing situation on the flex pipe. I'm not seeing exactly how the lightning is coupled to the stove (the vent hood, proximity of stove to lightning bolt, CCC's running over the roof???), but dissipating a high frequency transient with a length of 12 or 14 AWG is hopeless. Even 10 AWG is no good.
Here is a concrete example based on a quick lab test I just did (can't spend too much time on the extracurriculars, but I wanted to post with some hard data). I set up a transient generator for transient with a 2 uS rise and a 10 uS fall to half-crest with an open circuit voltage of 2.5KV and a short circuit current of 500 amps. This may sound like a hefty surge, but it is actually a surge that can be carried, without opening, by a typical 1.25A slo-blo fuse. I discharged this transient into a 12 foot length of 10 AWG wire. I have set up a transient waveform monitor (essentially a high voltage scope probe) to measure the voltage at a point half way along this wire to ground. What I find is about a 800 V bump lasting appr 3 uS at the point 6 ft away from ground (earth and generator return). When I increase the generator current to 2000 amps of potential short circuit current, this bump grows to about 1200 volts and stretches to about 12 uS total. (scope JPGs avail upon request)
This 2000 amp surge is what is coupled to ground from a single 4-wire telecom circuit (clamped to earth) that is expected to survive during GR-1089/NEBS testing. This explains why telecomm equipment manufacturers say to ground equipment using a 6 AWG. Ignore this at your customer's peril (had to throw that in since I know there are some who think the instructions to ground telco stuff with #6 is a bunch of hogwash. It isn't.)
But to the point of the flex pipe... Given the relatively low magnitude of my test, it is easy to see how a really big surge trying to discharge to ground through a much longer run of 14 AWG would develop an even larger potential at the flex pipe. The potential could be, and apparently is, great enough to puncture the jacket and then discharge to plumbing, other conduit, or other grounded objects that are not already carrying the transient current and are therefor still at the true ground potential (as they are not involved in any ground potential rise caused by high freq current through inductive wiring)
My 2 cents.