zeyadkhaled / cloudek-hackathor

Web Vulnerabilities Incidents Monitoring Service using Machine Learning. Awarded 2nd place at Hackathor.

ClouDek

Web Vulnerabilities Incidents Monitoring Service using Machine Learning

Description

Using ML models with 99.5% accuracy - %99.2 F1 Score, we are able to detect different web attack variants like (XSS,SQLi,CSRF,Open Redirect,etc..). The developer of a certain website could add our JS code, that routes all query params and submitted forms to our core Microservice that parses this data and then communicates to the ML microservice to get confidence results and then through a secure websockets connection send those results to a dashboard that has informative and attractive Widgets and Charts and Incoming Incidents alerts written in React and Redux.

Screens

Features

• Using ML models with 99.5% accuracy - %99.2 F1 Score
• Microservice architecture design
• Scalable
• Easy to deploy
• Test different attack variants (XSS,CSRF,SQL,Bruteforce)
• Setup in under 1 min
• Attractive & informative dashboard
• Secure communication protocols (SSL,WSS + Encryption)

Technologies used

• Python
• Tornado
• Scipy Kit
• Numpy
• Asyncio
• React
• Redux
• JQuery
• Digital Ocean

Author

• Zeyad Khaled
• Hikmet Demir
• Abdurahman Atakishiyev
• Ping Cheng