IPC-4552 requires 3-6um nickel, but does not appear to distinguish between rigid FR4 PCBs and flexible PCBs.

A flexible interconnect PCB using ENIG with ZIF end connectors has been shown to be prone to track cracking near the ZIF ends. The supplier has suggested a lower thickness of nickel 1-3um rather than the IPC spec to would be beneficial to resolve this and tests on samples seem to bear this out. 

Does anyone have any experience of such a suggested modification of ENIG spec and that thinner nickel improves the resistance to stress-cracking for flexi tracks? 

In this specific application as there are no component soldering requirements so the thinner nickel under the gold does not lead to solderability concerns, but what would be the likely consequence on other flexi/flex-rigid designs with flexi ZIF connector ends and SMT components if a 1-3um thickness was imposed more generally? Increased likelihood of black-pad type defects?