Guardian v1.0 — Deterministic Governance Infrastructure for AI Agents

Guardian introduces a deterministic governance layer between AI agents and execution.

Core capabilities:

• Policy as Code
• Deterministic ALLOW / DENY / ESCALATE decisions
• Evidence ledger for every decision
• Replay verification for auditability

Guardian is designed to act as the governance control plane for autonomous systems.

This first release includes the core governance engine, policy evaluation, evidence recording, and replay verification examples.