In this blog post, we present our approach for uncovering vulnerabilities by combining LLM reasoning with static analysis. By layering an LLM on top of CodeQL, we significantly reduce the...