EVA - EVA is an AI-assisted penetration testing agent th

项目说明

!E

<di="center">

⫻ 𝝣.𝗩.𝝠

⮡ Exploit Agent

<br>

Autonomous offensi AI for guiding pentest processes

![Stars]()
![Watchers]()
![Forks]()
![Repo Views]()


!GitHub issues
!GitHub pull requests
!GitHub contributors
!GitHub last commit

</div>

<br> <br>

<di="center">

![Checkout my other project :> NekoCLI!]()

</div>

---

𝝺 O

EVA is an AI penetration testing agent that guides users through complete pentest engagements with AI-powered attack strategy, autonomous command generation, and real-time analysis based on outputs. The goal is not to replace the pentest professional but to guide and assist and provide faster results.

Main funcionalities

🜂 Intelligent Reasoning: Ad AI-dri and attack path identification depending on query.

ⵢ Automated Enumeration: Systematic target reconnaissance and information gathering based on provided target.

ꎈ Assessment: AI-powered identification and exploitation strategies, suggesting next steps for or OSINT.

⑇ Multiple AI Backends: Support for Ollama, OpenAI GPT, Anthropic, Gemini, G4F.dev and custom API endpoints

ㄖ Session Management: Persistent sessions and chats

⑅ Interacti: Real-time command execution and analysis of output in multi-stage.

---

ⵢ E & Pentest Process Flow

---

<details>
<summary><h2>➤ Quick Start</h2></summary>

🍎 Installation

#### Ollama for local endpoint (required for local models and e database)

#### pip installation

#### E installation

⬢ Configuring EVA.

When starting EVA, it will automatically handle:

✅ API key setup (According to Model)

✅ Ollama model download (Default set as whiterabitv2, feel free to change to any other desired model)

✅ Session directory creation

✅ Dependencies installation

<strong> If you wish to modify endpoints, ollama models, API Keys or configure E run: </strong>

📁 Directory Structure of EVA

ꀬ Where to change E

<strong> Will display the following configuration: </strong>

</details>

<details>
<summary><h2>🖴 Usage Guide</h2></summary>

Initialization

1. Select Session: Choose existing session or create new one
2. Choose AI Backend:
- Ollama (Recommended): Local AI with WhiteRabbit-Neo model
- GPT-5: OpenAI's latest model (requires API key)
- G4F: Uses g4f.de with models running GPT5.2, feel free to change model used.
- Anthropic: Claude API backend (requires API key)
- Gemini: Google Gemini API backend (requires API key)
- Custom API: Your own API endpoint if desired
3. In the input field of chat, type in your request or what you need assistance with for EVA to help you!

> USER > i need help with a CTF machine, ip is 10.10.16.81

After making a request, commands will be provided and the pentest workflow will start, use commands below as reference.

| Command | Description |
|---------|-------------|
| /exit / /quit | Exit EVA and sa |
| /model | Change AI backend |
| /rename | Rename the current session |
| /search <query> or search <query> | Run exploit/ intel search inside current chat session and feed results into next analysis |
| /report | Generates a PDF/HTML report with latest findings on session |
| /map | Generates a html file with attack surface map of session |
| /menu | Return to session menu |
| R | Run suggested command |
| S | Skip command |
| A | Ask for next step |
| Q | Quit session |

ㄖ Example of chat session

> demonstration video.

!Demo Usage

</details>

<details>
<summary><h2>Ξ AI Backends</h2></summary>

🦙 Ollama (Recommended)

Model: ALIENTELLIGENCE/whiterabbitv2" (best one for OffSec)

- ✅ Complete offline operation
- ✅ No API costs
- ✅ Pri
- ❌ Higher CPU/GPU usage, recommended for machines above 8GB+ VRAM/RAM
- ❌ Heavier model, ~9.8gb model

⬡ OpenAI GPT

Models: GPT-5, GPT-4.1 (fallback)

About:

- ✅ Faster reasoning
- ✅ Extensi base
- ✅ Continuous updates
- ❌ Paid, requires apikey

ᛃ G4F.dev

Models: GPT-5-1

About:

- ✅ Updated information in real-time (usually)
- ✅ Quick responses
- ❌ Might be unstable or down sometimes, low stability.

⟅ Custom API

Endpoint: Configurable API to use your own as you wish. Please run e to set API handler and payload

About:

- ✅ Custom model integration
- ✅ Modifiable as you wish

✶ Anthropic

Model: Configurable via ANTHROPIC_MODEL

About:

- ✅ Strong reasoning quality
- ✅ Stable API
- ❌ Requires ANTHROPIC_API_KEY

✦ Gemini

Model: Configurable via GEMINI_MODEL

About:

- ✅ Fast response latency
- ✅ Native JSON output mode and better parsing, usually provides best results
- ❌ Requires GEMINI_API_KEY

</details>

<details>
<summary><h2>⑇ Roadmap</h2></summary>

[x] ⬢ OpenAI integration: Integrated OpenAI into EVA

[x] ⬢ G4F.DEV: Added G4F endpoints to free GPT5 usage.

[x] ⬢ Custom API: Add custom endpoint besides ollama and OpenAI

[x] ⬢ Automated Reporting: Concise HTML report generation (+ optional PDF )

[x] ⬢ C Integration: Real-time data

[x] ⬢ Attack Maps: Interacti diagrams such as connections or such, like Kerberos domains and AD devices.

[ ] ⬡ Cloud Integration: AWS/GCP deployment ready

[ ] ⬡ Web Interface: Browser-based E

</details>

<details>
<summary><h2>⨹ Legal Notice</h2></summary>

🚨 IMPORTANT

This tool is for allowed en only!

#### ✅ APPROVED USE CASES
> CTF (Capture The Flag) competitions <br>
> Authorized penetration testing <br>
> Security research and laboratory en <br>
> Systems you own or ha permission to test <br>

#### 🚫 PROHIBITED USE
> Unauthorized access to any system <br>
> Illegal or malicious acti <br>
> Production systems without explicit authorization <br>
> Networks you do not own or control

⚠️ DISCLAIMER

</details>

<details>
<summary><h2>⫻ License</h2></summary>

MIT License

</details>

<di="center">

❤️ Support

### if you enjoy the project and want to support future de:

![Star on GitHub]()
![Follow on GitHub]()
<br>

<a href=' target='_blank'><img height='36' style='border:0px;height:36px;' src=' border='0' alt='Buy Me a Coffee at ko-fi.com' /></a>
<br>
<strong>Hack the world. Byte by Byte.</strong> ⛛ <br>
𝝺𝗿𝗰𝗮𝗻𝗴𝗲𝗹𝗼 @ 2026

[[ꋧ]](#-𝝣𝗩𝝠)

</div>

---

*⚠️ Remember: With great power comes great responsibility. Use this tool ethically and legally.*

项目信息

开发语言：Python

Star 数：530

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EVA is an AI-assisted penetration testing agent that enhances offensive security workflows by providing structured attack guidance, contextual analysis, and multi-backend AI integration.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。