Experimental studies of protein pattern formation (both in vivo and in vitro) have stimulated new interest in the dynamics of reaction-diffusion systems. However, a comprehensive theoretical understanding of the dynamics of such highly nonlinear, spatially extended systems is still missing. In contrast, for dynamical systems described by ordinary differential equations, major insights can be gained from the geometric analysis of their phase-space structure, and this approach has proven invaluable in shaping our intuition about nonlinear dynamics. Here we show how such an analysis can be generalized to mass-conserving reaction-diffusion (McRD) systems, which are the generic case for intracellular protein pattern-forming systems.We present a comprehensive theory for two-component McRD systems, which serve as paradigmatic minimal systems that encapsulate the core principles and concepts of our framework. The key insight is that shifting local equilibria -controlled by the local mass density -give rise to concentration gradients that drive diffusive redistribution of total density. Thereby local chemical equilibria serve as a 'supporting framework' (referred to as scaffold here) for patterns during their entire dynamics.We show that the reactive nullcline (line of chemical equilibria) and the flux-balance subspace, where the diffusive fluxes are balanced, are the geometric objects that facilitate a full characterization of the reaction-diffusion dynamics in the phase plane of the reaction kinetics. (i) We demonstrate that the lateral instability underlying pattern formation in reaction-diffusion systems ('Turing instability') is a mass-redistribution instability (MRI). The mechanism underlying the emergence of MRIs is a feedback loop between shifting local equilibria and mass redistribution due to gradients induced by the shifting equilibria (mass redistribution cascade). The onset of an MRI is characterized by a simple geometric criterion based on the slope of the reactive nullcline. Whether the onset of MRI is sub-or supercritical is determined by the local curvature of the nullcline. (ii) Beyond the weakly nonlinear analysis, we find that the characteristics and bifurcations of stationary patterns can be graphically determined by a flux-balance construction on the reactive nullcline (akin to the 'common tangent construction' employed to find binodal and spinodal lines for phase separation at thermal equilibrium). (iii) By dissecting the patterns into spatial regions that exhibit characteristic signatures, we provide a complete quantitative characterization of all pattern types in two-component reaction-diffusion systems with mass conservation. Remarkably, our approach unambiguously reveals that excitability is inextricably linked to regional lateral instabilities and can be easily read off the nullcline shape.Taken together, these results demonstrate that phase-space geometry provides the essential elements of a comprehensive theory of pattern formation far beyond the linear regime -the effects of nonlinearities o...