Docs

Policy Evaluation

Evaluate applicable policies and determine the outcome of an AI invocation.

Curated Asserted, validated against implementation facts ai-control.policy-evaluation

Features

Priority-based evaluation
Allow, deny, warn, and audit effects
Subject matching
Target matching
Conditions
Rate-limit and token-budget conditions
Policy bundle evaluation

Provenance

This page is derived from the following implementation sources.

  • data-plane/modules/policy

Last verified at commit b61836e.