The intrinsic behavioral variability in resistive switching devices (also known as "memristors" or "ReRAM devices") can be a reliability limiting factor or an opportunity for applications where randomness of resistance switching is essential, such as hardware security and stochastic computing. The realistic assessment of ReRAM-based circuits & systems towards practical exploitation requires variability-aware ReRAM modeling. In this context, here we present a versatile, circuit-level implementation strategy to incorporate cycle-to-cycle (C2C) variability to the ReRAM model parameters in SPICE simulations. We evaluated the proposed approach with threshold-based models of a voltage-controlled bipolar ReRAM device and managed to reproduce the main features observed in experimental curves for different pulsed voltage inputs. With key upgrades, compared to previous approaches found in the literature, our strategy enables the enhancement of any ReRAM device model towards the exploration of new ways to make the most of the C2C ReRAM variability, and to test the robustness of any designed circuits & systems against ReRAM variability.