Overview
Org-wide policy for which auth routes and agents members use.
Proliferate's agent policy is how an organization admin states which authentication routes and which coding agents the team should be using, without touching anyone's individual setup by hand.
It lives on the Organization settings screen, visible only to owners and admins.
Agent policy
Organization settings, Agent policy section: two checklists (Allowed routes, Allowed harnesses) plus a Conflicts table listing members whose current selection falls outside the policy.
What it covers
- Source tiers: the three ways a member's agent can authenticate model calls (native sign-in, a raw API key, or Proliferate's managed gateway), and which coding agents the policy applies to.
- Enforcement: what actually happens when a member's setup doesn't match the policy today, and what requires a paid plan to change.
The short version
An admin checks which routes and which agents are allowed. Anyone on the team whose current selection falls outside that allow-list shows up in a Conflicts table on the same screen, by name, agent, and surface (cloud or local). Nothing is blocked automatically; the policy is a way to see and follow up on drift, not a hard gate.
Related
- Members & groups for who holds the admin role that can edit this policy.
- Guardrails for the other controls that shape what an agent can reach and do.