KG-based pre-populated dataset strategy (June 25 Agentic AI meeting)

From the June 25 Agentic AI meeting (raised by Jason): design a KG-based pre-populated dataset strategy that creates reusable AI-ready datasets after EDA, stores them as dataset nodes or references, and allows agents to fetch them directly instead of regenerating raw-data-derived datasets for each hypothesis. See the June 25 Agentic AI meeting notes for details.

Start with the June 25 Agentic AI meeting notes referenced in the issue, then inspect the existing dataset and knowledge-graph entry points. Define what reusable datasets store, how agents fetch them, and how completion will be verified without regenerating raw-data-derived datasets for each hypothesis.

Written by the indexing model from the issue text.