# MandateOS

> AI agent guardrails, approvals, and audit history for Codex, Cursor, Claude Code, OpenClaw, and MCP-based workflows.

MandateOS is an open-core system. The public repo contains the open-source trust layer: SDKs, the MCP server, installer CLIs, starter policy bundles, docs, and the homepage. The managed control plane handles hosted approvals, workspace operations, retained audit history, and customer administration.

## Primary

- Homepage: https://getmandateos.com/
- GitHub repo: https://github.com/robinsjovoll/mandate-os
- Open-source boundary: https://github.com/robinsjovoll/mandate-os/blob/main/OPEN_SOURCE_BOUNDARY.md
- Latest release: https://github.com/robinsjovoll/mandate-os/releases

## Public packages

- @mandate-os/sdk: https://www.npmjs.com/package/@mandate-os/sdk
- @mandate-os/mcp: https://www.npmjs.com/package/@mandate-os/mcp
- @mandate-os/openclaw: https://www.npmjs.com/package/@mandate-os/openclaw

## Core concepts

- Mandates define allowed tools, budgets, thresholds, approvals, and escalation policy.
- Agents are evaluated before sensitive actions continue.
- Higher-risk actions can require explicit approval or execution grants.
- If the runtime is unavailable, guarded actions stop instead of silently bypassing policy.
- Receipts and audit history preserve what was approved and what happened.

## Integrations

- Cursor
- Codex
- Claude Code
- OpenClaw
- Custom MCP-powered workflows