What problem does it solve? Analysts reasoning about retail source data often guess at row grain, join keys, and aggregation rules, producing metrics that cannot be reconciled. This Skill routes those questions through Seshat's curated SQL knowledge base so reasoning stays grounded in documented facts instead of assumptions. ## Core Features & Use Cases - Routed Knowledge Lookup: Opens the bi-sql-knowledge INDEX and selects only the route relevant to the current question, avoiding context bloat. - Grain and Join Grounding: Enforces explicit row-grain statements and forbids inferring missing source facts during join and aggregation reasoning. - Validation and Reconciliation Guidance: Provides documented rules for validation, reconciliation, and transformation decisions without executing any database. - Use Case: While mapping a messy retail extract into a medallion warehouse, ask how to aggregate line-item sales to daily store totals; the Skill follows the aggregation route and returns grounded rules with stop conditions. ## Quick Start Ask the assistant to use the bi-sql-knowledge skill to determine the correct grain and join keys for aggregating retail sales lines into daily store totals.
