Skip to content

AI-Powered SOC Threat Hunting Platform | Sysmon + Python Detection Engine + Machine Learning (Isolation Forest) + VirusTotal Enrichment + Flask Dashboard

License

Notifications You must be signed in to change notification settings

nikhil-p-00/ai-soc-threat-hunting-platform

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🌐 AI-Powered Threat Hunting & SOC Automation Platform A Mini-SIEM + SOAR + Threat Intelligence + AI Anomaly Detection System 🚀 Overview

This project is a fully functional AI-Powered Threat Hunting & SOC Automation Platform, combining:

Sysmon Log Collection

Rule-based Detection Engine

Machine Learning (Isolation Forest)

Threat Intelligence (VirusTotal Integration)

SOC Dashboard (Flask + Tailwind)

Automated Alerting & Enrichment

It simulates how modern SOC platforms (like Splunk, Sentinel, Elastic & CrowdStrike) work internally — but built entirely using Python, Sysmon, SQLite, and AI models.

This project is designed to showcase skills in:

🧠 Skills Demonstrated

  • ✔ Threat Hunting
  • ✔ Detection Engineering
  • ✔ SOC Automation
  • ✔ Python for Security
  • ✔ Log Parsing
  • ✔ AI/ML for Cybersecurity
  • ✔ Threat Intelligence Integration
  • ✔ Dashboard Design

🧩 Architecture Diagram

Windows Host (Sysmon) | v Log Collector (EVTX) | v SQLite Database | +-------------------+ | | v v Rules Engine AI Engine (Python Rules) (Isolation Forest) | | +---------+---------+ | v Threat Intelligence (VirusTotal) | v SOC Dashboard (Flask UI)

✨ Features

🔍 1. Sysmon Log Collection

Reads Sysmon EVTX logs directly from C:\Windows\System32\winevt\Logs\Microsoft-Windows-Sysmon%4Operational.evtx

Automatically extracts:

Event ID

Image

CommandLine

Source/Destination IP

Raw XML

🛡️ 2. Detection Engine

Includes multiple types of detections:

✔ Suspicious PowerShell ✔ Encoded commands (-enc) ✔ Unknown/rare process execution ✔ External IP connections ✔ Behavior-based alerts

Built to mimic EDR-style detection logic.

🤖 3. AI Anomaly Detection

Uses an Isolation Forest model to detect:

Unusual process behavior

Abnormal command-line patterns

Rare system events

Each alert is tagged:

NORMAL ANOMALY

With a numerical score.

🌍 4. Threat Intelligence (VirusTotal)

For any event containing an IP:

Automatically queries VirusTotal

Adds malicious/suspicious count

Adds direct VT permalink

Example enriched alert:

PowerShell suspicious command
ANOMALY score=0.053
VT: malicious=5 suspicious=2 |

📊 5. SOC Dashboard

A modern Flask dashboard that shows:

Live alerts

Severity levels

Rule hits

Anomaly scores

VT enrichment tags

📁 Project Structure AI-SOC-Project/ │ ├── collector/ │ └── sysmon_collector.py │ ├── detection/ │ ├── rules.py │ └── sysmon_rules.py │ ├── ai/ │ ├── train_anomaly.py │ └── model.joblib │ ├── dashboard/ │ ├── app.py │ ├── templates/ │ └── static/ │ ├── tools/ │ ├── vt_lookup.py │ ├── add_vt_column.py │ └── insert_test.py │ ├── Sysmon/ │ └── sysmonconfig-export.xml │ ├── events.db └── README.md

🧪 How to Run 1️⃣ Start Sysmon Collector python collector/sysmon_collector.py

2️⃣ Run Detection Engine python detection/sysmon_rules.py

3️⃣ Train AI Model (optional) python ai/train_anomaly.py

4️⃣ Start Dashboard python dashboard/app.py

Then open:

👉 http://127.0.0.1:5000

🔥 Example Alerts ALERT [HIGH] PowerShell suspicious command: id=42 cmd=powershell -enc Y2FsYy5leGU= ANOMALY=NORMAL score=-0.011 VT: malicious=4 suspicious=1 https://virustotal.com/...

🧰 Technologies Used

Python 3

Flask

SQLite

Sysmon

scikit-learn

TailwindCSS

VirusTotal API

Regular Expressions

Windows Event Log Parsing

🏆 Why This Project Is Valuable

This project demonstrates real-world skills required for:

SOC Analyst

Detection Engineer

Threat Hunter

Security Automation Engineer

MDR Analyst

Recruiters love seeing:

Real detection logic

AI anomaly detection

Threat intel enrichment

Dashboard + backend

This is not a basic project — it is portfolio-grade.

📌 Disclaimer

This platform is for educational and defensive research purposes only. Do not use it to monitor systems without authorization.

🎉 Final Note

This project reflects modern cybersecurity engineering practices and shows strong capability in:

Writing production-grade security code

Building defensive tools

Understanding attacker behavior

Designing SOC automation logic

About

AI-Powered SOC Threat Hunting Platform | Sysmon + Python Detection Engine + Machine Learning (Isolation Forest) + VirusTotal Enrichment + Flask Dashboard

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published