sdlc-agent-05-mapping-document

What problem does it solve? Data engineering teams need consistent, validated source-to-target mapping documents during the SDLC, but producing them manually from metadata discovery results is slow and error-prone. This Skill acts as Agent 5 in a 15-agent Databricks pipeline, turning a Metadata Discovery Package into a contract-validated mapping_document.json artifact stored in the lakehouse. ## Core Features & Use Cases - Mapping Generation: Produces source-to-target table/column mappings, transformation logic, lookup/join logic, SCD strategy, and incremental vs full-load decisions per target entity. - Contract Validation: Returns JSON matching a strict output schema, with one repair retry on validation failure and routing to human review when clarification is needed. - Lakehouse Integration: Writes artifacts to /Volumes/{catalog}/state/artifacts/{ticket}/ and logs runs to state.artifacts and state.agent_runs for full auditability. - Use Case: After the Metadata Discovery Agent profiles source systems for a Jira ticket, this agent generates the mapping document that downstream Data Model and PySpark Development agents consume, gated by human approval. ## Quick Start Hand this agent the Metadata Discovery Package for ticket PROJ-123 and ask it to produce the source-to-target mapping document artifact.