Few static analysis tools exist for detecting security-critical issues in .NET, and fewer yet which can be deployed at build-time. In the first part of the talk, we discuss the work required to support taint analysis with Infer#, one of the first scalable tools for statically detecting .NET security issues. We outline themes that we hope will prove useful for other frontend developers wishing to improve taint support. In the second part of this talk, we consider the significant recent progress in Large Language Models and the opportunities they present when used for program repair and defect summarization, specifically in conjunction with Infer.