Coach — User brain

Competency map

Editable self-report; living profile, not a test.

AI usage and prompting
Strong on iteration; still building prompt habit discipline.
Context environments and knowledge curation
Built the Butternut operating layer; knows where context rots.
System architecture and workflow design
Coding and technical implementation
Ships via AI-assisted build — does not hand-write code.
Commercial forecasting and planning
Data quality and evidence
Team leadership and change
Domain-specific work (field sales ops)

Profile-based calibration

From your competency scores: how Clive's pace and Pam's sensitivity adjust.

Signal	Clive	Pam
Strong commercial / domain judgement	Trust domain claims more; ask sharper operational questions	Focus on narrative risk, stakeholder reaction, overconfidence

Coach Whit: prompt coaching

Seeded tips keyed to your weakest domains. Calibrated enablement.

Ask for evidence, not enthusiasm
Replace 'is this a good idea?' with 'what assumption would Pam challenge first?' — you get challenge without skipping the human gate.
What is the weakest assumption in this rollout plan, and what evidence is missing?
Scope the build lane explicitly
Tell Doc whether you want a proposal, a package, or a Cursor dispatch brief. Narrow scope keeps Composer work bounded.
Doc, prepare a build brief for a brain health page, seeded demo data only, no live system writes.

Champion signals (read-only)

Optional context for coaching, not a performance score.

Engagement level
High — actively shaping the brain and reviewing agent answers
Set by Self (founder champion)
Technical adeptness
Commercial-first; relies on Doc lane for implementation
Set by Self (founder champion)