Pwno is a stateful system for autonomous binary research and exploitation, specifically designed to empower LLM agents with advanced debugging and control capabilities.

How does Pwno enable LLM agents for exploitation?

It runs GDB + pwndbg in an isolated environment, exposing stateful debugging, exploit I/O, and helper tooling over MCP for programmatic control by agentic coding clients.

What are Pwno's core features?

Key features include stateful GDB/pwndbg debugger sessions, deterministic execution control via GDB/MI, fast context snapshots, interactive exploit-driver workflows, and multi-session support.

Is Pwno suitable for security testing and vulnerability research?

Yes, Pwno provides essential tools for security testing, vulnerability research, and reverse engineering binaries by offering detailed stateful debugging and exploit development capabilities designed for automated agents.

Pwno

Name: Pwno
Author: pwno-io

bypwno-io

•

232

•

Developer Tools

Data Science & ML

Security & Testing

Provides a stateful system for autonomous binary research and exploitation, specifically designed for LLM agents.

Pwno is a powerful, stateful system engineered for autonomous binary research and exploitation, primarily serving LLM agents. It integrates GDB and pwndbg within an isolated environment, exposing comprehensive debugging capabilities, exploit I/O, and helper tooling over the MCP protocol. Key features include deterministic execution control, rapid context snapshots of registers, stack, disassembly, and source, alongside interactive exploit-driver workflows. The system also supports multi-session parallel operations and offers workspace automation helpers for managing commands, processes, Python scripts, repositories, and reverse engineering tools like RetDec.

Key Features

01Interactive exploit-driver workflows with pwncli

02Stateful debugger sessions with GDB and pwndbg

03Deterministic execution control via GDB/MI

04Fast context snapshots for registers, stack, and disassembly

05Multi-session support for parallel tasks

06230 GitHub stars

Use Cases

01Orchestrating parallel exploit development workflows

02Automated binary research and analysis for LLM agents

03Developing and testing exploits in an isolated environment