What problem does it solve? Designing a Malloy semantic model requires deciding which tables to model, at what grain, and which dimensions and measures to define. Doing this from schema alone leads to wrong grains, arbitrary bucket thresholds, and untested business assumptions. ## Core Features & Use Cases - Source plan proposals: Recommends base sources per table with grain and primary key, plus computed sources for grain mismatches and repeated aggregation patterns, including dependency mapping and deferred sources. - Data-driven field definitions: Proposes renames, dimensions, and measures per source, each backed by query results such as distributions, distinct values, and percentile-based bucket boundaries. - Business logic clarification: Flags decisions that cannot be made from data alone, such as which of two date columns is canonical, with specific evidence. - Use Case: Given an orders database, the skill inspects the existing model, queries value distributions, and proposes a confirmed source architecture and field set ready to write into the .malloy model. ## Quick Start Ask the assistant to propose a Malloy source plan and field definitions for the orders and customers tables, with every proposal backed by querying the data.
