Agent setup

The point of Specline is that you don't carry the methodology and you don't paste the canon — the agent learns it from the tool, and you bring the judgment. Two pieces make that automatic.

1. The repo carries a brief

Drop a CLAUDE.md at the repo root. Claude Code reads it on open, so any agent you start there is briefed without you saying a word. A ready template lives at specline/templates/specline-CLAUDE.md; it tells the agent it's a Specline repo, how to shape and migrate plans with you, and to use the doctor tools.

2. doctor is available as a tool

With the MCP server registered at user scope, the agent discovers doctor_spec / doctor_rules / doctor_check in every repo. The brief plus the tools mean the agent self-onboards.

The day-to-day flow

1. Open a fresh agent chat in the repo (with CLAUDE.md present).
2. Say it in your own words:
   you →

   Let's shape a new feature: <your idea>.
   — or —
   Migrate this plan into Specline with me: <the file>.

3. The agent pulls the canon (doctor_spec), asks you the shaping questions, and writes the spec folder — checking itself with doctor_check (author mode) as it goes.
4. You read the result in one sitting and approve it, or send it back. That's the ratify gate — your call, not the machine's.

Without Claude Code

The same works for any agent with shell access: have it run specline spec to load the canon and specline doctor --mode author --format json to check its work. MCP just makes discovery automatic.