Security Domain Specific Languages (DSLs) bring structure and and tractable reasoning to security properties of systems. The power of Security DSLs can introduce assurance at the level of Automated Reasoning (AR) to unverified systems without requiring developers to adopt tools like formal methods. In other words, Security DSLs can make high-assurance development accessible to developers without the capacity to apply AR themselves. However, taking advantage of the full power of Security DSLs still requires security experts who understand the context that their systems operate in. In this talk, I will show how we can apply AR techniques to automatically and on-demand infer the context that systems operate in to generate Security DSL programs to automatically evaluate security and compliance properties. Combining Security DSLs and AR creates a flywheel that lets us assess the security of systems more quickly and comprehensively than is possible with either alone.