Security problems are created throughout a system’s development lifecycle, but often are only discovered after it is put into use. These problems become particularly worrying in safety-critical systems, the failure of which can cause injuries or even death. Reasoning about security at the architectural level is helpful to build systems that are more resistant to attack, but can also be challenging, as much of the existing security research is focused on low-level design aspects. In this talk, I’ll discuss recent and ongoing work on architecture-level security using the SEI’s model-based development language and tooling.