In the future, webs of unmanned air and space vehicles will act together to robustly perform elaborate missions in uncertain and sometimes hostile environments. To achieve this robustness we go beyond current embedded programming languages, introducing a model-based programming language that enables autonomous vehicles to select and adapt coordinated mission plans on the fly. First, we present a variant of the Reactive Model-based Programming Language (RMPL), that allows for the expression of complex concurrent activites, metric time constraints and multiple contingencies. Second, we introduce the Temporal Planning Network (TPN), a simple, compact encoding of all possible executions of an RMPL program, that supports a model of program interpretation as fast temporal planning. Finally, we introduce an RMPL interpreter, called Kirk, which uses graph search on TPNs to find temporally consistent executions of RMPL programs.