The policy engine.
The model proposes; the policy engine decides.
What it is
A deterministic rule layer between your agent and the venue. Every action the model proposes is checked against your rules before it can execute. There is no model judgment inside the check, so the same inputs always give the same answer.
The rules you set
- Loss limits. A cap on what an agent may lose. A proposal that would breach it is rejected.
- Asset allowlists. The markets an agent may trade. Anything off the list is rejected.
- Approval rules. The actions that need your explicit sign-off before they execute.
- Payload-bound approvals. An approval binds to the exact proposed action. If the action changes, it is a new proposal.
When something is rejected
It does not execute, and the check is part of the turn record, so you can read why in the debrief.