Mormukut976/DeepGuard-AI
GitHub: Mormukut976/DeepGuard-AI
这是一个集成了钓鱼邮件检测、日志网络异常分析及大模型安全防护的AI驱动企业级网络安全平台。
Stars: 0 | Forks: 0
# 🛡️ DeepGuard AI - 企业网络安全平台
[](https://python.org)
[](https://fastapi.tiangolo.com)
[](https://streamlit.io)
[](https://attack.mitre.org)
[](https://docker.com)
[](LICENSE)
## 🚀 核心功能
| 功能 | 描述 | 状态 |
|---------|-------------|--------|
| 📧 **网络钓鱼检测** | 基于 DistilBERT 的邮件分析 | ✅ |
| 🔐 **LLM 安全** | Prompt 注入检测（符合 Zelis AI 标准） | ✅ |
| 📊 **日志异常** | 使用 Isolation Forest 进行异常检测 | ✅ |
| 🌐 **网络安全** | 实时数据包捕获 | ✅ |
| 🎯 **MITRE ATT&CK** | 完整的 v14.0 映射 | ✅ |
| 🛡️ **对抗鲁棒性** | FGSM/PGD 攻击模拟 | ✅ |
| 📤 **SIEM 集成** | Microsoft Sentinel + Slack/Email | ✅ |
| 🔧 **CI/CD 安全** | Bandit + Safety + Trivy | ✅ |
## 🏗️ 架构
DeepGuard AI/
├── 📧 网络钓鱼检测器 (DistilBERT)
├── 📊 日志异常检测器 (Isolation Forest)
├── 🌐 网络异常检测器 (Isolation Forest)
├── 🔌 FastAPI 后端
└── 💻 Streamlit 前端
text
## 🛠️ 安装说明
1. **克隆仓库**
```
git clone
cd DeepGuard-AI
Install Dependencies
bash
pip install -r requirements.txt
Train Models
bash
python train_models.py
Start API Server
bash
python run_api.py
Start Frontend (New Terminal)
bash
streamlit run frontend/app.py
Access System
API Docs: http://localhost:8000/docs
Dashboard: http://localhost:8501
📊 API Endpoints
POST /analyze/phishing - Analyze emails for phishing
POST /analyze/logs - Detect anomalies in system logs
POST /analyze/network - Monitor network traffic
POST /analyze/comprehensive - Complete security scan
GET /system/status - System health check
🧠 ML Models
Phishing Detection: Fine-tuned DistilBERT model
Log Analysis: Isolation Forest for anomaly detection
Network Analysis: Isolation Forest for traffic patterns
🎯 Usage Examples
Phishing Detection
python
import requests
response = requests.post("http://localhost:8000/analyze/phishing",
json={"emails": ["You won $1000! Click here..."]}
)
print(response.json())
Comprehensive Scan
python
payload = {
"emails": [...],
"logs": [...],
"network_traffic": [...]
}
response = requests.post("http://localhost:8000/analyze/comprehensive", json=payload)
📈 Performance
Phishing Detection Accuracy: ~90%
Log Anomaly Detection: ~85%
Network Threat Detection: ~88%
Response Time: < 2 seconds
🔧 Development
Backend: FastAPI + Python
Frontend: Streamlit
ML: PyTorch, Scikit-learn, Transformers
Data: Pandas, NumPy
```
标签：AMSI绕过, Apex, AV绕过, CI/CD安全, Cloudflare, DevSecOps, DistilBERT, FastAPI, FGSM, HTTP/HTTPS抓包, Kubernetes, Llama, Microsoft Sentinel, MITRE ATT&CK, NLP, PGD, Python, SIEM集成, Streamlit, 上游代理, 人工智能, 企业安全, 凭据扫描, 大模型安全, 威胁检测, 孤立森林, 安全运营, 对抗样本, 开源安全工具, 异常检测, 态势感知, 扫描框架, 无后门, 机器学习, 深度学习, 用户模式Hook绕过, 系统调用监控, 网络安全, 网络流量分析, 网络资产管理, 网络钓鱼检测, 访问控制, 请求拦截, 逆向工具, 逆向工程平台, 邮件分析, 隐私保护