HackSynth

HackSynth is a sophisticated LLM-based agent designed for autonomous penetration testing using a dual-module architecture comprising a Planner and a Summarizer. Its primary use case is to conduct security assessments, and it is benchmarked against two extensive CTF-based datasets derived from PicoCTF and OverTheWire, featuring 200 diverse challenges. Notable features include its iterative command generation and feedback processing capabilities, enabling comprehensive evaluation of LLM penetration testing agents.