Headless CRM / Project skills
CRM Setup
- name
- crm-setup
- description
- Use when the CRM tables described in /crm/SETUP.md do not exist yet, the user asks to set up or initialize the CRM, or list_tables shows a table from the setup doc is missing.
Goal
Turn the schema specified in /crm/SETUP.md into real tables, once, without improvising a different shape.
Steps
- Call list_tables. If all five tables already exist, stop and say so — do not recreate or "fix" a table a human may have adjusted on purpose.
- Read /crm/SETUP.md. It is the specification; you are executing it, not redesigning it. If a column looks wrong, raise it before creating anything.
- Ask only what setup genuinely needs: what is sold, to whom, and who owns the pipeline. Do not interview the user about fields.
- Confirm the plan in a few lines — five tables, one line each — then create them with create_data_table exactly as specified. The built-in setup-data-table skill covers the mechanics if you need them.
- Write what you learned into the business context section of /SUBTERRANEAN.md.
- Report what was created, then give the activation step verbatim: open a session with Pipeline Steward and say "activate".
Do not
- Do not create tables the setup doc does not list. A CRM that needs eleven tables on day one is a CRM nobody will keep current.
- Do not seed example rows. An empty pipeline is honest.
- Do not configure another agent's triggers or heartbeat. That is step 4's job, in Pipeline Steward's own session.