Build a golden-question benchmark for data agents: frozen Iceberg tags, execution-based scoring with Arrow, and CI gates.
A plan for when a data agent gives a wrong number: severity levels, containment, Iceberg time travel reproduction, and blast radius queries.
Serve many customers from one lakehouse without leaking between them: isolation models, Polaris grants, and tenant-scoped agent state.
Ship prompts, tools, models, and metric changes without breaking answers: agent manifests, shadow mode, canaries, and rollback.
