// Advanced Cyber Security WINNIX DEFENDER

Cross-platform malware detection powered by LightGBM. Scan Windows PE & Linux ELF binaries statically — zero execution, full SHAP transparency on every single verdict.

Get Setup Files Get Thesis PDF

Windows

Linux Debian / Kali

Offline Protection

⚡LightGBM Classification 🔬54 Binary Features 🛡️Windows PE Support 🐧Linux ELF Support 🧠SHAP Explainability 🌐VirusTotal Integration 📡AbuseIPDB Checks 🔒Quarantine Manager 📊Live Dashboard 📁Batch Directory Scan 💾SQLite Diary 📄PDF / CSV Export ⚡LightGBM Classification 🔬54 Binary Features 🛡️Windows PE Support 🐧Linux ELF Support 🧠SHAP Explainability 🌐VirusTotal Integration 📡AbuseIPDB Checks 🔒Quarantine Manager 📊Live Dashboard 📁Batch Directory Scan 💾SQLite Diary 📄PDF / CSV Export

Core System Capabilities

Built to Detect & Protect

WinNix Defender signature dependency ko khatam karta hai aur explainable machine learning se threats ko spot karta hai.

🤖

AI Engine & Feature Profiling

System LightGBM classifiers ko custom-engineered PE aur ELF datasets par train karta hai. Yeh static metrics se accuracy secure karta hai aur SHAP transparency show karta hai.

✓ SHAP Interpretability displays feature priority
✓ Section Entropy Analysis identifies modifications
✓ LightGBM swift training maintains fast local runtime
✓ Static API Calls profiling tracks payload traces

LightGBM Model Engine

🔬

Deep Binary Parsing Core

Windows Portable Executable (PE) aur Linux Linkable Format (ELF) files ko LIEF library se parse karke 54 dimensional structural metrics ko extract kiya jata hai.

✓ High-speed extraction for both .exe/.dll and .so/.bin files
✓ Structural byte-level magic number type confirmation
✓ Zero execution policy ensures malicious code never runs
✓ Auto quarantine mechanism isolates suspected payloads

LIEF Framework core

Binary Features

<300^ms

Avg Scan Speed

Supported Formats

93.7^%

ML Scan Accuracy

Knowledge Base & Thesis

Get Certified Academic Docs

WinNix Defender Software Engineering final year program ka detailed research project hai jo Hazara University ke strict standards par design kiya gya hai.

01

Introduction & Problem Statement (Ch 1)

Traditional signatures aur local machine learning detection techniques ka fundamental comparison.
03

System Methodology & Feature Engineering (Ch 3)

Extracting 54 key structural details like section entropy, raw data size aur import traces.
04

Testing & Performance Validation (Ch 4)

Accuracy checks, black box testing, QThread responsiveness aur security audits parameters.
05

SHAP Explainability Results (Ch 5)

Features values contribution charts aur predictive algorithm decision charts behavior evaluation.