Skip to content

LinkDefender AI — An explainable, ethical AI system that detects and prevents LinkedIn-style phishing attacks by analyzing recruiter messages, job posts, and profile credibility using NLP, anomaly detection, and human-interpretable XAI insights.

Notifications You must be signed in to change notification settings

Slothcoder310/TheGoodWill-Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

🔐 LinkDefender AI — Detect. Explain. Educate.

LinkDefender AI is an AI-driven cybersecurity solution built to identify and explain social-engineered phishing attempts on LinkedIn and professional networks.

With the rise of fake recruiter messages, cloned company profiles, and malicious job posts, traditional spam filters fail to detect these context-based attacks.

Our system integrates Natural Language Processing (BERT fine-tuned), Anomaly Detection, and Explainable AI (SHAP/LIME) to:

Flag suspicious messages and fake job posts in real time.

Highlight why a message is risky using interpretable explanations.

Educate users on common red-flag patterns and social-engineering cues.

🧩 Key Features:

Explainable AI: Understand the decision behind every detection.

Dataset-driven Training: Uses curated phishing data (linkedin_phish_1000.csv, fake_job_postings.csv, and verified_online.csv).

Ethical & Transparent: GDPR-compliant preprocessing and anonymized data usage.

Modular Architecture: NLP pipeline → Feature extraction → Model scoring → Explainability layer.

💡 Goal: Empower professionals to recognize, report, and resist trust-based phishing attacks — turning awareness into the first line of defense

About

LinkDefender AI — An explainable, ethical AI system that detects and prevents LinkedIn-style phishing attacks by analyzing recruiter messages, job posts, and profile credibility using NLP, anomaly detection, and human-interpretable XAI insights.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages