ctf-agent - Autonomous CTF solver that races multiple AI model

项目说明

CTF Agent

Autonomous CTF (Capture The Flag) solver that races multiple AI models against challenges in parallel. Built in a weekend, we used it to solve all 52/52 challenges and win 1st place at BSidesSF 2026 CTF.

Built by Veria Labs, founded by members of .;,;. (smiley), the #1 US CTF team on CTFTime in 2024 and 2025. We build AI agents that find and exploit real security for large enterprises.

Results

| Competition | Challenges Solved | Result |
|-------------|:-:|--------|
| BSidesSF 2026 | 52/52 (100%) | 1st place ($1,500) |

The agent sol across all categories — pwn, re, forensics, web, and misc.

How It Works

A coordinator LLM manages the competition while sol attack indi challenges. Each swarm runs multiple models simultaneously — the first to find the flag wins.

Each solver runs in an isolated Docker container with CTF tools pre-installed. Solvers never give up — they keep trying different approaches until the flag is found.

Quick Start

Coordinator Backends

Sol

Default model lineup (configurable in backend/models.py):

| Model | Provider | Notes |
|-------|----------|-------|
| Claude Opus 4.6 (medium) | Claude SDK | Balanced speed/quality |
| Claude Opus 4.6 (max) | Claude SDK | Deep reasoning |
| GPT-5.4 | Codex | Best o solver |
| GPT-5.4-mini | Codex | Fast, good for easy challenges |
| GPT-5.3-codex | Codex | Reasoning model (xhigh effort) |

Sandbox Tooling

Each solver gets an isolated Docker container pre-loaded with CTF tools:

| Category | Tools |
|----------|-------|
| Binary | radare2, GDB, objdump, binwalk, strings, readelf |
| Pwn | pwntools, ROPgadget, angr, unicorn, capstone |
| Crypto | SageMath, RsaCtfTool, z3, gmpy2, pycryptodome, cado-nfs |
| Forensics | , Sleuthkit (mmls/fls/icat), foremost, exiftool |
| Stego | steghide, stegseek, zsteg, ImageMagick, tesseract OCR |
| Web | curl, nmap, Python requests, flask |
| Misc | ffmpeg, sox, Pillow, numpy, scipy, PyTorch, podman |

Features

Configuration

Copy .en to .env and fill in your keys:

All settings can also be passed as en or CLI flags.

Requirements

Acknowledgements

项目信息

开发语言：Python

Star 数：828

技术要点

该项目属于CTF思路方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕Autonomous CTF solver that races multiple AI models in parallel. 1st place BSidesSF 2026.展开，提供了可运行的实现，对理解相关安全机制、工具原理与攻防技术具有较高的研究价值。

从技术实现角度看，此类项目的核心在于对底层原理的深入理解与工程化落地。开发者在研究过程中可以结合官方文档与社区讨论，梳理项目的初始化流程、核心模块职责与数据流转，逐步掌握其中的技术细节与设计思路。

使用与学习建议

1. 下载源码压缩包，解压到本地环境。

2. 根据项目类型准备对应的运行环境与依赖。

3. 按照 README 中的步骤完成配置与部署。

4. 在隔离的测试环境中运行验证，避免影响生产系统。

免责声明

本资源来源于开源社区，仅供安全研究、技术学习与合法授权范围内的攻防演练使用，请勿用于任何非法或侵害他人权益的用途。使用本项目产生的一切后果由使用者自行承担。