Framework for Cloud Data Security Using Agentic AI

Naseer, R and Srujan, K M and Deepthi, A S and Divyashree, C H and Goutham, M (2025) Framework for Cloud Data Security Using Agentic AI. International Journal of Science and Research Archive, 15 (1). pp. 1730-1735. ISSN 2582-8185

Cloud platforms are ever more vulnerable to advanced cyber threats, which demand intelligent and self-reliant security systems. We introduce CloudShield, a prototype Agentic AI system for mimicking live cloud data protection in Azure platforms. The system mimics round-the-clock log collection in Azure-type protocols, employs the Isolation Forest algorithm to identify outliers, and responds automatically to attacks such as brute-force attacks and malware. Logs are locally stored in a SQLite database, encrypted for secure storage, and can be deployed entirely self-contained without dependencies. CloudShield includes an interactive real-time dashboard for threat visualization and system analysis. Its agent-based, modular architecture supports scalability, automation, and high detection rates—making it a strong experimental model for current cloud security research.