Smart India Hackathon 2026

AI-powered IPsec VPN security analysis

An AI-powered framework designed to generate IPsec VPN testbeds, capture encrypted network traffic, identify VPN protocols and security configurations, analyze metadata, and automatically assess cybersecurity risk.

PS ID
26160
Theme
Blockchain & Cybersecurity
Category
Software
College
ALIET
Cyberpunk mascot: a hooded, glowing lime reaper with a pixel-dissolving scythe
IKE stateIKE_SA_INIT detected
TrafficESP, AES-256-GCM
RiskLow, 92 / 100
AI engine online
VPN testbed active
Network monitor running
Security engine scanning
Dataset connected
Ollama AI ready

Five modules take a VPN from setup to verdict

Each configuration is built, exercised, captured, identified and assessed by the same five stages.

01

VPN testbed generation

Builds a repeatable IPsec lab in tunnel or transport mode, with the cipher, DH group and PFS setting you choose.

  • Tunnel mode
  • Transport mode
  • AES-128
  • AES-256
  • AES-GCM
  • AES-CBC + HMAC
  • DH groups
  • PFS
  • IPv4 / IPv6
02

Traffic generation

Pushes realistic application traffic through the tunnel so captures look like real use.

  • VoIP
  • Messaging
  • E-mail
  • Web browsing
  • ICMP
  • Video streaming
  • TCP / UDP
03

Traffic capture

Records the key exchange and the encrypted packets into PCAP files.

  • Wireshark
  • tcpdump
  • Custom utilities
  • PCAP
  • IKE negotiation
  • ESP packets
  • AH packets
04

AI identification

Works out what the VPN is doing from the capture alone, without reading any payload.

  • IPsec protocol
  • IKE version
  • Tunnel / transport
  • Encryption
  • Auth algorithm
  • Key exchange
  • Traffic inference
05

Security assessment

Scores the configuration against known-good practice and explains every deduction.

  • Crypto strength
  • Config compliance
  • SA parameters
  • Key lifetime
  • Replay protection
  • Forward secrecy
  • Metadata

Follow one configuration through the system

Eleven stages carry a researcher's settings all the way to a report. Pick a stage to see what goes in and what comes out.

Trained in the cloud, run locally

Models are trained in Google Colab, exported, then served on your own machine through Ollama for offline, privacy-preserving inference.

Cloud

  1. 01Google Colab
  2. 02Dataset preparation
  3. 03Model training
  4. 04Model evaluation
  5. 05Model export

Local

  1. 06Ollama local AI
  2. 07Cyberpunk security engine
  • Protocol classification
  • Traffic classification
  • Security configuration analysis
  • Metadata inference
  • Risk prediction
  • Security recommendations
Any OS. One App.

How to install CYBERPUNK

Get the complete AI-powered IPsec VPN security analysis framework running locally on your machine.

  1. STEP 01 Download cyberpunk.zip

    Get the core application package containing the frontend dashboard, Python backend, and analysis scripts.

    Download cyberpunk.zip
  2. STEP 02 Install Ollama

    Our local AI analysis engine relies on Ollama. Download and install it for your OS (Windows, macOS, or Linux) to enable offline, privacy-preserving inference.

  3. STEP 03 Run setup.py

    Extract the ZIP file, open your terminal in the extracted folder, and run the setup script. This installs required libraries and connects Cyberpunk AI with Ollama.

    $ python setup.py

ANY OS ONE APP

Cross-platform compatible

Fully containerized and architecture agnostic. Run the entire security dashboard and AI analysis pipeline natively on Windows, macOS, or any Linux distribution seamlessly.

Technology stack

AI & machine learning
Google ColabPythonOllamaMachine learning models
VPN & networking
IPsecstrongSwanIKEv1 / IKEv2ESP / AH
Network analysis
WiresharktcpdumpPCAP analysisPython Scapy
System
LinuxDockerVirtual network namespaces
Hardware
ESP32Wi-Fi network monitoringIoT communication
Backend
Python server
Frontend
HTML / CSSJavaScriptReact

Automated security reports

Open any report to preview it. Generating a full report requires installing and running the application locally.

Executive report

High-level security summary for management.

Technical report

Detailed VPN configuration and packet analysis.

Risk score

Automated cybersecurity risk evaluation.

Threat matrix

Visual representation of identified risks.

AI confidence score

Confidence level of AI predictions.

Dataset report

Training and testing dataset information.

Team CYBERPUNK

Shaik Haroon Rasheed

N. Hemanth

Minas Venkat Siddu

G. Venkata Nagalakshmi

Bathala Sarayu

Chukkla Revanth

Security teamCollege: ALIETSmart India Hackathon 2026

Mission deliverables

    Report