What problem does it solve? Databricks AI/BI (Lakeview) dashboards use a unique serialized JSON structure where wrong widget versions, mismatched field names, or untested SQL cause broken dashboards. This Skill enforces a strict workflow that validates every query through the Databricks CLI before deployment. ## Core Features & Use Cases - Guided dashboard creation workflow: Discover table schemas, probe data with aggregate SQL, plan widget/filter structure, then deploy with databricks lakeview create using the correct flag and JSON field placement. - Full widget specification references: Versioned specs for counters, tables, bar/line/pie charts, maps, forecast-line with AI_FORECAST, pivot, histogram, sankey, heatmap, funnel, box, waterfall, and filter widgets. - Theming and layout rules: 12-column grid layout, uiSettings theme palettes, semantic color pinning, and cardinality limits for readable charts. - Use Case: Ask the agent to build a support operations dashboard on your Databricks tables; it explores the schema, tests each dataset query via CLI, generates the dashboard JSON, deploys it, and publishes it with a shareable link. ## Quick Start Create a Databricks AI/BI dashboard showing weekly case volume, resolution time KPIs, and a region filter from the catalog.schema.support_cases table.