We present an architecture that enables run-time verification with executable models of behaviour. Our uniform modelling paradigm is logic-labelled finite-state machines (LLFSMs). Behaviours are constructed by parameterizable, loadable, and suspendable LLFSMs executed in predictable sequential schedules, but they are also verified at run-time by LLFSMs as well. Our architecture enables runtime verification (to monitor the quality of software during execution) as well as set up, tear down, and enforcement of quality behaviour during runtime. The LLFSMs models are executable and efficient because they are compiled (not interpreted). The LLFSMs can be derived from requirement engineering approaches such as behaviour trees, and also validated using test-driven development. However, in situations where software evolves incorporating elements of adaptive systems or machine learning, the software in execution may have never existed during development. We demonstrate the features of the architecture with illustrative case studies from robotics and embedded systems.