zhaoxuya520/reverse-skill

AI agents often lack domain knowledge about which reverse engineering or pentesting tool to use for a given task, leading to guesswork and repeated mistakes. This framework solves that by providing a structured routing system that maps task types (APK, ELF, JS, PCAP, CTF) to proven playbooks, automatically checks available tools, and bootstraps missing toolchains on demand. It also captures and reuses experience through a self-evolving knowledge base, preventing the same errors from recurring.

Developers can build AI-assisted security analysis workflows that handle APK/Android reverse engineering, iOS/mobile analysis, frontend JS encryption cracking, binary analysis (ELF/PE), PCAP analysis, and CTF challenges. The ceiling is high: with the routing and toolchain bootstrapping, an AI agent can autonomously execute complex multi-step security assessments, generate evidence-based reports, and maintain a field journal of findings. Real-world use cases include automated malware analysis, vulnerability research, authorized penetration testing, and security education.