Analyze a table

Use when: "what does this table do?", "explain the {table} workflow", "walk me through this table", "what's set up here?" — a question about the table's structure and the workflow encoded in its columns, not about any single record.

A Clay table is a column DAG: source and input basic columns are roots; basic formula columns and action (enrichment) columns depend on the {{f_xxx}} tokens in their settings. Analyzing the table = reconstructing that graph, ordering it into stages, and narrating what it does. Cheap and read-only: ~3 commands (metadata + columns + a row sample).

Finding the table: resolve it as in the tables entry-point skill ("Finding a table").

1. Frame the table — clay tables get

(The output is discriminated on type — shape in clay tables get --help.) Read off the frame the workflow sits in:

clay tables get <tableId> | jq .

• type — "archive" means this is an archive companion (its parentTableId points at the regular table); analyze the parent instead, the archive has no workflow of its own.
• rowCount — current size.
• archive — non-null means processed rows flow out to an archive, indexed on archive.searchableFieldFormula.