Build the practice
brain first.
Every agent you deploy is only as good as what it knows about you and your practice. The Practice Brain is that knowledge, structured and committed before your first agent runs.
One brain. Every agent in your harness reads from it.
Most AI tools start with a blank slate every session. The Practice Brain is a living model of your practice that every agent in ClineOS reads from, writes to, and updates over time.
Two methods. One complete picture of your practice.
A webform captures the surface. The Practice Brain goes deeper — resolving ambiguity, locking decisions, and building shared language before any agent acts on your behalf.
Practice Grill: align before you deploy.
DrClaw interviews you — one dependency at a time — until every key term is precise and every assumption is named. It doesn't stop at a surface answer.
Practice Pre-Mortem: assume failure. Build from there.
Assume the agent fails. Work backwards. Every hard-to-reverse decision gets documented as an ADR before DrClaw or DrCline touches a single task.
Brain Commit: your context becomes the source of truth.
Your practice model is structured and committed to ClineOS. Every agent in your harness reads from the same foundation — no separate onboarding, no repeated context.
Agents deploy. All reading from the same brain.
DrClaw runs your operations. DrCline joins your clinical team. AppBuilder ships your patient site. Everything reads from the same brain — and every session makes it sharper.
Built from what's already in your head.
Most of this you know. None of it is written down anywhere agents can read. The Practice Grill surfaces it, names it, and structures it so your agents can act without asking you to repeat yourself.
Clinical Identity
- Specialty and sub-specialty focus
- Clinical decision frameworks you trust
- How you communicate diagnoses
- Specialty protocols you've built over time
Practice Model
- Care delivery format (DPC, fee-for-service, hybrid)
- Revenue cycle and payor mix
- Visit cadence and access model
- Referral relationships and preferences
Patient Population
- Panel composition and size
- Chronic condition mix
- Demographics and risk profiles
- Geographic and socioeconomic context
Operations and Values
- Staff roles and escalation paths
- Communication preferences
- What you will not compromise on
- How you want patients to feel
It gets smarter every time you use it.
The Practice Brain isn't a form you fill out once. Every session adds context. Every agent interaction surfaces something new. The brain versions itself, and the practice running on session 50 is fundamentally different from the one on session one.
Agents stop starting from scratch
DrClaw and DrCline already know your panel, protocols, and preferred phrasing. Context is committed, not re-entered.
Each session updates the brain
New decisions, refined protocols, and corrected assumptions write back to CONTEXT.md. The brain is always current, not frozen at onboarding.
Context score compounds over time
The brain tracks completeness. As gaps close and edge cases resolve, your agents operate with higher confidence and narrower margins for error.
Institutional knowledge stays in the building
Your clinical judgment doesn't walk out the door when a staff member leaves. It lives in the brain, versioned and owned by you.
New agents inherit everything
Every module you add to ClineOS reads from the same foundation. No separate onboarding. No repeated setup.
Yours to own, not rent
The brain lives inside ClineOS. Not a vendor's training set, not a shared model. A practice asset that appreciates.
The brain that runs your practice in month six looks nothing like the one in month one.
Every session closes a gap. Every corrected assumption improves the model. The context score isn't an onboarding metric — it's a live measure of how well your practice is understood by the agents running it.
Your practice is already a system.
Give it a brain.
The brain that runs your practice gets sharper with every session. Start building it before your first agent goes live.