The first agentic reverse engineer (github.com)

Kong, an innovative reverse engineering tool, has emerged as the first fully autonomous agentic system capable of deobfuscating and analyzing binaries with near-human efficiency. Utilizing a robust NSA-grade framework, Kong automates the traditionally labor-intensive process of reverse engineering, where simple tasks can take hours or even days. It uses a five-phase pipeline, orchestrating analysis from function triage to semantic synthesis, leveraging the power of large language models (LLMs) like Anthropic’s Claude and OpenAI’s GPT-4o. Notable features include in-process Ghidra integration, call-graph-ordered analysis, and a unique agentic deobfuscation pipeline that identifies and neutralizes obfuscation techniques frequently employed in binary security. This development is significant for the AI/ML community as it demonstrates a breakthrough in automating code analysis, which has traditionally required significant human expertise and insight. By combining rich context windows, systematic function analysis, and automatic naming recovery, Kong not only speeds up the reverse engineering process but also enhances accuracy, recovering function names and structures that are typically lost in stripped binaries. With applications ranging from malware analysis to software security, Kong could transform the landscape of binary analysis, making advanced reverse engineering more accessible and efficient for developers and security professionals alike.