Developed a dataset-level quality check for the databrickslabs/dqx repository, introducing the has_no_aggr_outliers feature to detect anomalies in time-series aggregates. Leveraging PySpark and Python, the solution applies a stateless rolling-window sigma approach to dynamically flag outliers based on historical data trends, enhancing data quality governance for analytics workflows. The work encompassed comprehensive unit and integration testing, performance validation against production-like datasets, and thorough documentation updates, including usage demos. By addressing both technical robustness and usability, the contribution reduces the risk of undetected outliers impacting metrics and supports more reliable decision-making in data-driven environments.