What problem does it solve? In a multi-agent Databricks SDLC pipeline, someone must turn the upstream Data Model and DDL into production SQL artifacts (views, procedures, reconciliation queries) with validated syntax and a machine-readable output contract. This Skill defines Agent 8 of 15, which performs that SQL generation step deterministically and hands a validated artifact to the next agent. ## Core Features & Use Cases - SQL artifact generation: Drafts Gold-layer views, procedures, and reconciliation queries via the Databricks Foundation Model API, then validates syntax with EXPLAIN on a SQL warehouse. - Genie Agents routing: Routes natural-language business questions to the Genie Agents Conversation API instead of hand-writing ad-hoc SQL, recording the generated SQL for traceability. - Contract-disciplined output: Returns JSON matching a strict output contract, writes artifacts to /Volumes/{catalog}/state/artifacts/{ticket}/08_sql-agent/, and logs runs to state.agent_runs. - Use Case: A ticket arrives with a data model from Agent 6; this agent produces validated sql/*.sql files, upserts artifact metadata, and signals the Data Quality Agent to continue. ## Quick Start Given the data model and DDL artifacts for ticket 1234, generate and validate the Gold-layer SQL views and write them to the ticket's 08_sql-agent artifact folder.
