ShadowGuard AI

• 

  front page

• 

  app.py

• 

  showing risk

• 

  full screen

Inspiration

In today’s digital world, every individual leaves behind a “digital shadow” through social media activity, online accounts, and browsing behavior. I was inspired by the growing number of cyberattacks, identity theft cases, and data breaches happening globally.

I wanted to build a system that does not just react to cyber threats but predicts them before they happen.

Problem Statement

Most cybersecurity tools work after a threat occurs. There is very little focus on predictive risk analysis for individuals.

If we denote user risk as:

[ Risk = f(Exposure, Password Strength, Public Data Leakage) ]

then reducing exposure and strengthening security factors can significantly lower the probability of cyber attacks.

Solution – ShadowGuard AI

ShadowGuard AI analyzes publicly available digital footprints and calculates a personalized cyber risk score.

Key Features:

• Public data exposure analysis
• Weak password detection
• Dark web breach monitoring simulation
• Predictive cyber risk scoring
• Personalized security recommendations

How We Built It

• Frontend: React.js for interactive dashboard
  
• Backend: Python (Flask)
  
• Machine Learning: Scikit-learn for risk prediction model
  
• Database: Firebase / MongoDB
  

We trained a classification model to predict whether a user falls into Low, Medium, or High cyber risk categories.

What I Learned

• How predictive analytics can be applied in cybersecurity
  
• Feature engineering for risk assessment
  
• Real-time data handling and dashboard visualization
  
• Importance of ethical AI and responsible data use
  

Challenges Faced

• Collecting meaningful sample data
  
• Avoiding privacy violations
  
• Optimizing model accuracy within hackathon time limits
  
• Ensuring real-time performance
  

Future Scope

• Integration with real-time dark web APIs
  
• Browser extension for live monitoring
  
• AI-powered auto security patch suggestions
  

ShadowGuard AI aims to move cybersecurity from reactive defense to proactive prediction.

Built With

• css
• flask
• github
• html
• javascript
• python
• scikit-learn
• sqlite