Dislodged nitride particles are hard enough to promote galling of ANY steel, when the substrate becomes deformed the relatively thin surface treatment can act as an abrasive, equal hardness materials will always cut each other with enough force applied. Once the surface treatment is compromised friction welding of the soft bolt and hard receiver steel becomes possible.
Your annealing comments give me the impression the nitride is a gas or salt bath process as opposed to a plasma or PVD duplex process.
I would agree that the steel seems unsuitable for task, from the pictures I have seen the surface finishes involved may also be unsuitable for task. I would be interested to see some pictures of the receivers involved. The reality of needing a certain amount of clearance to facilitate sliding of the components means there is always some unpredictable concentration of forces due to imperfect operation inherent to humans.