Extratropical cyclones (ETCs) are a primary driver of extreme precipitation and near-surface wind speed across the mid-latitudes, with major societal and economic consequences over large parts of North America. Yet assessing the future impact of ETCs through the quantification of changes in their associated extremes remains a major challenge, owing to the coarse resolution of global models and the complex interplay of dynamical and thermodynamic mechanisms. Three regional climate simulations from the CRCM6-GEM5 model, covering the NA-CORDEX domain at 12 km grid spacing, and driven by boundary conditions from the EC-Earth3-Veg, MPI-ESM1-2-HR, and MIROC6 CMIP6 atmospheric-ocean general circulation models (AOGCMs), alongside an ERA5-driven simulation, are used to (i) assess the representation of ETC-associated extreme precipitation and  winds relative to the ERA5 reanalysis and (ii) project their late-century changes under the SSP3-7.0 scenario. To this end, a novel extreme exceedance framework is used to isolate changes into three physically interpretable drivers: the intensity of ETC-associated extremes, the background extreme threshold, and the frequency of co-occurrence of ETCs and extremes. CRCM6-GEM5 systematically produces stronger hourly extremes than the ERA5 reanalysis, for both precipitation and wind speed. The driving boundary conditions constitute the dominant source of inter-simulation spread in accumulated extreme exceedances, primarily through their control on the frequency of extreme ETC events. Under future climate forcing, all three simulations robustly project an amplification of ETC-associated extreme precipitation over northeastern North America and the adjacent northwestern North Atlantic, driven by a concurrent intensification in individual ETCs and increased occurrence. Elsewhere, large inter-simulation spread in the occurrence response precludes robust conclusions on projected changes. Applying the framework to the most extreme precipitation-producing systems impacting northeastern North America in the historical and future climates, we found a consistent intensification across multiple exceedance metrics in all three simulations. Future changes in ETC-associated extreme wind speed, however, show little coherence across simulations and are dominated by a redistribution of occurrence rather than any systematic intensification. These results underscore the critical role of large-scale boundary conditions in shaping regional projections.