Set up auth, session, and profile config for clay-cli before running any command - use when a Clay command fails with "session expired", when starting fresh in a new checkout, or when configuring workspace/folder/workbook IDs for the first time.
Discover Clay actions and manage the local action catalog - use when searching for an available action/integration, checking catalog coverage against the strict proof registry, classifying action safety, or generating action templates.
Extract Clay table/workbook data to local files - use when you need a full-fidelity workbook export, a template-shaped snapshot of a table, or raw paginated row dumps for offline processing.
Run a full Clay playbook end to end - intake, plan, prepare, sample, verify, scale - use when a plain-language request should be routed to one of the 12 packaged playbooks and operated through to a safe scale decision, not just a single ad-hoc command.
Build Clay list-building sources - company-to-people waterfalls, source preview/import, and webhook intake - use when adding a new list-building source to a table or converting a companies table into a people search.
Build and modify Clay tables, views, fields, and records with clay-v2.js - use when creating a new table/workbook, adding or editing columns (text/formula/http-api/use-ai/action), managing views, or writing declarative table specs.
Read-only exploration of Clay tables, workbooks, sources, and workspace metadata - use when you need to understand a table's shape, check run status, list workspaces/tables, or pull credit usage without changing anything.
Run a small sample through Clay enrichment columns and verify the results - use when you need to fire an action on rows (credit-consuming), watch a run to completion, or read back and score a table before deciding whether to scale.