Hydrogen embrittlement poses a significant obstacle to the integrity of common structural alloys, impacting the safety and reliability of components for advanced technologies. Despite decades of research, the origins of hydrogen-assisted failure are not well understood, largely because we lack a fundamental understanding of how hydrogen environments affect the mesoscale localization of plastic deformation and damage initiation and propagation. Here, we examine the role of specific grain boundary configurations to understand the mechanistic origins of hydrogen-assisted failure to identify mitigation strategies for hydrogen embrittlement through microstructure design. Based on observations of hydrogen-enhanced stress localization, we hypothesize that high modulus mismatches across low-energy grain boundaries serve as preferential initiation sites for hydrogen-assisted failure. In this talk, we present the results of an extensive study of bicrystal configurations of austenitic stainless steel to assess this hypothesis. We use analytical tools and finite element simulations of crystal plasticity to identify conditions that promote the localization of plastic slip with and without the effects of absorbed hydrogen. We then build a model to quantify these observations and map the severity of plastic localization to a handful of a priori observable quantities, including mismatch of elastic modulus, thus providing an intuitive understanding of the mesoscale origins of hydrogen embrittlement. Sandia National Laboratories is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.