airt - AIRT — A free, open-source AI Red Teaming course w

项目说明

AIRT — AI Red Team Academy

A free, open-source course co offensi testing of AI systems — from prompt injection to supply chain attacks. 60+ hours of content with hands-on Docker labs.

🌐 View the course →

Modules

| # | Module | Topics |
|---|--------|--------|
| 1 | Foundations of AI Red Teaming | MITRE ATLAS, OWASP LLM Top 10, threat modeling |
| 2 | Prompt Injection Attacks | Direct/indirect injection, jailbreaks, filter bypasses |
| 3 | RAG Exploitation & Database Attacks | Knowledge base poisoning, embedding attacks |
| 4 | Multi-Agent System Exploitation | Agent hijacking, tool abuse, memory poisoning |
| 5 | AI Supply Chain & Infrastructure Attacks | Model backdoors, pickle exploits, dependency attacks |
| 6 | Model Extraction & Inference Attacks | Model stealing, membership inference, side channels |
| 7 | Automated AI Red Teaming at Scale | garak, PyRIT, promptfoo, CI/CD integration |
| 8 | Post-Exploitation & Impact Analysis | Lateral mo, reporting, regulatory frameworks |

Hands-on Labs

Each module includes a Docker-based lab en. No cloud API keys needed — e runs locally via Ollama.

Quick Start

Prerequisites

License

Content is licensed under CC BY-SA 4.0. Code and lab files are licensed under MIT.

---

Built with Perplexity Computer.

项目信息

开发语言：HTML

Star 数：231

技术要点

该项目属于红蓝攻防方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕AIRT — A free, open-source AI Red Teaming course with 8 modules and hands-on Docker labs. Built with Perplexity Computer.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。