viorelyo / extwatcher

Cloud based framework for detecting malicious files using Machine Learning 🎓

ExtWatcher

Cloud based framework for detecting malicious files using Machine Learning.

Description

ExtWatcher evolved from the idea of Bachelor Thesis Project. At the moment it has support just for analyzing PDF files. But there are plans for integrating other ML models for analyzing different file formats.
By assembling more components into ExtWatcher, we came up with a more complex software system. We have developed:

1. Windows Service for detecting the downloaded files, blocking and uploading them to the Analyzing Framework. It can take the corresponding action on that blocked files based on the scanning result (unblock + keep / delete).
2. System Tray Application for catching the events thrown by Windows Service when a file is being scanned. The events are transformed into Windows System Notifications.
3. Windows Installer that wraps both the Windows Service and the System Tray App in order to automate the configuration and start of the application.
4. React Dashboard for visualizing the metadata generated after scanning files. As a bonus it offers a crossplatform solution that implies submitting the URL of a file and the Framework will download and analyze it automatically.

Overview

Built with

• C#, .NET
• Python Flask
• Scikit-learn
• ReactJS

Links

Thesis
Presentation

Demos

• Windows Service + System Tray App

• ExtWatcher Dashboard