What problem does it solve? Semantic models often reach production without structured review of grain, relationships, metrics, and business meaning, leading to broken dashboards and mistrusted metrics. This Skill enforces a human-approved release process for Exasol Semantic Views models, turning ad-hoc review into an auditable, gate-driven workflow. ## Core Features & Use Cases - Phased Review Workflow: Classifies each review as DISCOVERY, TECHNICAL_REVIEW, SEMANTIC_REVIEW, RELEASE_REVIEW, or POST_RELEASE_REVIEW, with explicit decision owners and blocked questions. - Evidence and Validation: Builds an evidence map from source comments, constraints, profiles, and samples; queries CURRENT_VALIDATION_ISSUES and requires a clean VALIDATE_MODEL result. - Approval Gates and Handoff Bundle: Enforces SCOPE, TECHNICAL, SEMANTIC, and RELEASE approvals, then produces a complete review bundle with metric cards, semantic diff, acceptance tests, and feedback backlog. - Use Case: Before publishing a new revenue metrics model, run this Skill to verify grain and join correctness, reconcile metrics against trusted SQL, collect domain-owner sign-off, and deliver a READY or NOT_READY decision package to the data engineer. ## Quick Start Use $exasol-semantic-reviewer to review the candidate semantic model 'sales_analytics', identify unresolved technical and semantic risks, and produce an approval-ready handoff for the data engineer.
