See It Work — Roles you author — not ones the platform hands you | S2 Vol 2 · The Primacy Cockpit

Most AI platforms ship you pre-built roles you adopt. This cockpit inverts that: roles are operator-defined reality — you author them. Each role's scope (what it can and can't do) is explicit and rendered on screen, and changing a role's powers is itself a gated act, not a silent toggle.

Adopt the platform's default roles and you've inherited its assumptions about how your operation works. Author your own and the cockpit reflects your rules — the powers your AI has are the ones you deliberately granted.

What this means for you

The roles are yours to author, their scopes are visible, and changes are governed. What this means for you: your AI agents have exactly the powers you deliberately gave them — no inherited platform defaults quietly granting more than you intended, and no role widening without your approval.

Each role is operator-authored with an explicit, visible scope:

Roles

who authors rolesyou — not the platform

scopeexplicit, on screen

changing a rolea gated act, not a setting

the cockpit reflectsyour rules

Operator-defined reality — the cockpit shows the powers you granted, not the vendor's prescriptions.

For the technical reader — the command, and how to verify it yourself

# one line · you do not need to run this
open atrium-standalone.html

open atrium-standalone.html
# -> roles you authored, each scope explicit on screen

Full step-by-step is in Appendix RX: Hands-On Demonstrations in the book.