CyberShield AI: A web-based cyber threat monitoring and risk assessment prototype

CyberShield AI is a prototype for organising cybersecurity event information and identifying potentially suspicious activity through a web interface and backend services. The project report describes a dashboard, threat-scanning workflow, database records and automated alert concepts. Its supplied code appendix provides a Flask REST service, SQLite storage, and a deterministic riskscoring function using failed-login counts and request counts. A separate Random Forest training script demonstrates a machine-learning extension using a small illustrative dataset; the supplied code does not establish that this trained model is used by the running detection endpoint. This paper documents the prototype architecture, describes the implemented scoring logic, identifies limitations of the available evidence, and proposes a reproducible evaluation plan. The work is presented as a student prototype rather than a validated production intrusion detection system.