Get Schema

Describe the organization's data schema. Returns every table the agent can target, including its semantic category (e.g. raw vs. curated), description, and full column definitions with types, nullability, uniqueness, and foreign-key targets. Each column carries two names: 'name' is what the read and write tools accept, and 'sql_column_name' is the physical column for SQL queries (foreign keys carry an _id suffix). The org's own tables are followed by the platform tables (core_experiment, core_instrument, and similar) that SQL queries can join against; those have no id or semantic category. This is the map a parser writes into: use it to plan which tables to populate and how rows link. Requires an API key with the `read` scope.