Artificial Intelligence1 min read

LLM Assisted Reverse Engineering

By · Published by Everything Blog

In short

The workflow directs artificial intelligence agents to analyze structural binary properties while continuously maintaining context through Markdown log files. Operators track the system's reasoning traces, actively intervening to redirect focus toward alternative paths like binary strings when standard decompilation methods stall. For intricate investigations, the setup employs a hypothesis debugging framework that utilizes structured CSV logging to curb repetitive guesswork. When baseline static analysis fails, agents are tasked with executing dynamic steps such as injecting logging probes or operating emulators.

Key points

  • Incorporating artificial intelligence agents into reverse engineering substantially accel…: Incorporating artificial intelligence agents into reverse engineering substantially accelerates the inspection of complex software across multiple platforms.
  • Dynamic capabilities like emulator usage, probe injection, and breakpoint management help…: Dynamic capabilities like emulator usage, probe injection, and breakpoint management help agents bypass traditional decompiler bottlenecks.
  • Structured CSV hypothesis tracking prevents models from looping through repetitive, unstr…: Structured CSV hypothesis tracking prevents models from looping through repetitive, unstructured guesses during deep analytical tasks.

Sources checked by our analysis pipeline: 1

I have been hacking both hardware and software for 2 decades. Previously manually. But in 2026 with the help of AIs the proccess is greatly sped up. Below I will describe my proccess briefly or how I do it.

Set Up

Tools which I mostly use are these:

TOOL | use |

---|---|

DeepSeek Harness + Deepseek v4.1 Flash | Thats the AI I am using. Occasionaly I use Codex with Luna or Terra but Deepseek is super performant, fast and cheap |

Ghidra + MCP bridge | The best in the market, free, disassembler and decompiler. With the MCP bridge agents can use it very well |

radare2 and kuna | CLI disassemblers and decompilers. Not as good as Ghidra but much faster and possibly better usable for AIs |

Process

What I basically do nowadays is just point the agent at the binary to be analysed and use my set of dedicated agent skills. Rest usually takes care of itself. But putting the skills besides, the flow is basically this.

- Find binary type, platform, build etc -…

Artificial Intelligence