A major goal of polymer science is to develop "smart" materials that sense specific chemical signals in complex environments and respond with predictable changes in their mechanical properties. Here, we describe a genetic toolbox of natural and engineered protein modules that can be rationally combined in manifold ways to create reversible self-assembling materials that vary in their composition, architecture, and mechanical properties. Using this toolbox, we produced several materials that reversibly self-assemble in the presence of Ca2+ and characterized these materials using particle-tracking microrheology. The properties of these materials could be predicted from the dilute solution behavior of their component modules, suggesting that this toolbox may be generally useful for creating new stimuli-sensitive materials.