AI Governance — Rego Policies for OPA

4 open-source Rego policies encoding AI system governance and authorization controls. Apache 2.0, Rego v1, ready to load into OPA or Enterprise OPA.

What this is

4 Rego policy files encoding AI system governance and authorization controls, part of the Rego Policy Libraries — a standalone, dependency-free collection of Open Policy Agent policies for published security and compliance standards. No orchestrator, no agent, no vendor runtime.

Load and query it

git clone https://github.com/ynotbhatc/rego_policy_libraries.git
opa eval -d rego_policy_libraries/governance/ai -f pretty 'data'

Or load into a running OPA:

for f in rego_policy_libraries/governance/ai/*.rego; do
  curl -X PUT --data-binary @"$f" "$OPA/v1/policies/$(basename $f .rego)"
done

Two ways to use these policies

Gating a change of state — evaluate a proposed action before it proceeds, and refuse it if it violates the standard. Measuring state — evaluate collected facts about an environment against the standard to produce compliance evidence. The same policy content serves both; only the integration differs.

Policy modules

action_classification.rego ai_governance.classification
ai_governance_complete.rego ai_governance
authorization.rego ai_governance.authorization
context_validation.rego ai_governance.context

Contributing