A machine learning–powered tool that helps security teams prioritize vulnerabilities and generate automated remediation guidance. The system analyzes CVE descriptions, predicts severity (Low, Medium, High, critical) using logistic regression with vector embeddings, and generates remediation steps using a fine-tuned T5 model.
Collects and processes CVE data from NVD feeds (10+ years).
Converts CVE descriptions into vector embeddings for classification.
Predicts vulnerability severity (Low, Medium, High).
Generates remediation steps with a fine-tuned T5 model.
Interactive Streamlit web app for real-time testing.
CVE Data: Collected from NVD feeds (past 10 years).
Remediation Data: Scraped from vendor advisories, reports, and knowledge bases.
Start the remediation chatbot service:
python remediationchatbot.py
Launch the Streamlit app:
streamlit run app.py
Open the local URL shown in your terminal (default: http://localhost:8501) to use the tool.
Upload a vulnerability scan report (e.g., an OpenVAS .csv file).
The system will automatically extract CVE descriptions and process them.
For each vulnerability, the tool will return:
Predicted Severity (Low / Medium / High / critical)
Suggested Remediation Steps