Safe and effective interactions between humans and complex systems represent a requirement for practically all of NASA's missions. The first-of-a-kind nature of such systems and the range of unpredictable events they are likely to be confronted with raises considerable design issues. To address this challenge we propose the development of an environment for Evolvable Work-practice Interfaces between Humans and Agents, which adapts activity theory to the situated nature of action. During Phase I we have developed a scenario for modeling interactions with Advanced Life Support Systems, we have modeled the associated human activities, and criteria through which activities and actions are employed. The experiments conducted using the BRAHMS work-practice simulation environment have demonstrated the feasibility of the approach and its potential to generate and capture critical interaction cases. During Phase II we will expand our scenario to include additional potential challenges, we will expand our model for describing work activities and work practices, and the way they operate during human system interaction. We also propose to expand the simulation capabilities of the BRAHMS environment to increase the insight of experimenters into their simulations, and we will provide criteria for evaluating the soundness and performance of the proposed model and implementations.