ScamShield AI Job Scam Detector

ScamShield is a browser extension that analyzes job postings for risk indicators using AI, helping job seekers make safer, more informed decisions before applying or responding to a recruitment offer. HOW IT WORKS When you open a job posting and click Analyze Now, the extension reads the visible text and runs it through a multi-layer AI engine. The result is a Trust Score from 0 to 100, a risk level classification, and a breakdown of specific risk patterns found in the posting text. All analysis is performed on the content of the posting — not on any individual, recruiter, or company. Postings are classified as Low Risk, Moderate Risk, High Risk, or Very High Risk. These are probabilistic, educational assessments intended to guide further due diligence — not definitive statements about any person or organisation. HOW TO USE 1. Install the extension 2. Navigate to any job posting 3. Click the ScamShield icon in your browser toolbar 4. Click Analyze Now A Trust Score and risk factor breakdown will appear within seconds. No account, login, or configuration is required. A Gemini AI key is included and works immediately after installation. You may also enter your own free Gemini key in the Settings panel at aistudio.google.com. WORKS EVERYWHERE ScamShield automatically analyzes job listings on major hiring platforms. You can also manually paste any job description — from an email, a message, or any website — for instant analysis. WHAT THE ANALYSIS CHECKS • Unrealistic compensation for entry-level or low-skill roles • Requests to communicate via third-party messaging apps • Requirements for upfront payment, registration fees, or training bonds • Suspicious or mismatched contact email domains • Domain age and DNS validity of mentioned websites • Missing company information or absence of qualification requirements • Duplicate posting signals across multiple platforms • Brand impersonation by unverified entities TECHNICAL DETAILS The analysis engine combines Gemini AI with ETFF-Net, a custom machine learning model trained on 17,880 real and fraudulent job postings. ETFF-Net uses a fine-tuned DistilRoBERTa transformer for semantic embedding, dual TF-IDF feature extraction, 14 structural metadata features, attention-based feature fusion, and a LightGBM classifier. The model achieves 98.7% accuracy and a ROC-AUC score of 0.990 on the held-out test set. PRIVACY ScamShield does not collect personal data. When a scan is triggered, the job posting text is sent securely over HTTPS to the Google Gemini API solely to perform the requested analysis. No data is retained on any server operated by this extension. Scan history is stored locally in your browser and can be cleared at any time from within the extension. Privacy Policy: https://github.com/naagasumukh8/Job_Verify_FYP/blob/main/PRIVACY.md DISCLAIMER ScamShield analyzes patterns in job posting text only. All risk assessments are probabilistic and educational. The extension does not make accusations against any individual, recruiter, or company. Always apply personal judgment before acting on any result.