Misconfiguration Detection Dataset for Cloud IAM and APIs

This dataset was developed as part of a cloud security research framework targeting detection of Identity and Access Management (IAM) and Application Programming Interface (API) misconfigurations in AWS environments. It was designed to simulate cloud infrastructure activity using synthetic AWS CloudTrail logs and IAM policy configurations, enriched with labelled data for supervised learning. The dataset supports advanced research in anomaly detection and cybersecurity, particularly focusing on over-permissive roles, unauthorized credential usage, and suspicious API activity. It includes structured log data, such as event names, role names, session attributes, timestamps, and misconfiguration flags which are captured across diverse simulated cloud scenarios . This structured dataset is applied to graph-based learning to detect the misconfigurations. With a modular structure and CSV formatting, the dataset integrates seamlessly into various machine learning models. This resource is ideal for academic and industry researchers working on cloud threat detection, IAM policy analysis, and secure serverless deployment.