Dynamic assessments in general, and game-based assessment (GBA) specifically, compel us to rethink prevailing language testing conceptualizations of context. Context has traditionally been portrayed with a cognitive orientation, which focuses on static abilities, ignores complex interactions, devalues the role of tasks in determining scores, and makes connections to learning potential difficult. "Ability-in language user-in context" lays the groundwork for a new framework for use in language testing and GBA by shifting the conversation of interaction and context to a level of entanglement not yet considered in the field. The paper makes connections to concepts in other disciplines (e.g., intensive and extensive context, reciprocal interaction, and co-construction) and problematizes relevant design, measurement, and validity considerations. We identify claims in the areas of generalizability, score comparability, and dimensionality and suggest areas for future research. We end the article by inviting language testers in general, and those in the Asian language testing community in particular, to engage in GBAs.