What problem does it solve? In a multi-agent Databricks SDLC pipeline, unit test results alone do not prove a pipeline works end-to-end. This Skill acts as Agent 12 of 15, executing integration, regression, and performance tests against a staging schema and producing a consolidated test report that downstream Documentation and Deployment agents consume. ## Core Features & Use Cases - End-to-end test execution: Runs integration, regression, and performance tests against a staging schema/cluster, validating source counts versus target counts and spot-checking transformed rows. - Contract-validated output: Returns JSON matching a strict output contract (status, artifact_path, citations, tool_calls, tokens) that the LangGraph orchestrator validates before handoff. - Lakehouse artifact persistence: Writes test_report.md to /Volumes/{catalog}/state/artifacts/{ticket}/12_testing-agent/ and upserts tracking rows into state.artifacts and state.agent_runs. - Use Case: After the Unit Test Agent finishes, the pipeline hands its results to this agent, which runs regression suites on staging, detects a row-count mismatch, and returns needs_clarification with specific questions instead of inventing results. ## Quick Start Hand this agent the Unit Test results for a ticket and ask it to run integration, regression, and performance tests on the staging schema and write the test report artifact.
