An Agentic Retrieval Framework for Autonomous Context-Aware Data Quality Assessment

The paper introduces a unified agentic-retrieval framework for autonomous context-aware data quality assessment, utilizing large language models to derive context-specific assessment strategies and generate executable validation logic through a multi-agent workflow. A key feature is the feasibility validation stage, which ensures that generated specifications are realistic and executable, thereby enhancing operational reliability and enabling iterative refinement. This framework addresses the challenges of static assessment methods, offering a scalable solution for practitioners in data analytics to ensure data quality in diverse usage scenarios.