MandateOS logoMandateOSGuardrails, approvals, and receipts for agent action
Menu
Sign In

Safety guardrails for AI agents

MandateOS gives you a controlled loop for agent safety: you define guardrails, MandateOS intercepts risky actions, escalates approvals, and keeps signed evidence of every run.

Try for free See how it works

60-second tour

Watch one approval and receipt happen end to end.

Installer, mandate, runtime decision, signed receipt, and audit verification — without leaving the workspace.

How it works

What gets written, checked, kept, and inspected.

What Gets WrittenLocal MCP + hook config

MandateOS writes host-scoped files such as `.codex/config.toml`, `.codex/hooks.json`, Cursor hooks, and Claude workspace settings.

What Gets CheckedMandate, tool, zone, approval path

Before risky actions continue, the runtime evaluates scope, budget, risk zone, and whether an operator decision is required.

What Operators KeepReceipts, approval events, audit chain

Teams can review signed receipts, execution grants, approval events, and retained audit history after the action completes.

What You Can InspectPublic repo, packages, installers

The trust layer is open source, so teams can inspect the host integration path instead of relying on a black box.

See it running

Centralized visibility and control.

The MandateOS control plane brings full visibility to your agents' hooked actions. Easily draft custom governing rules, deploy smart security presets with the Mandate Composer, and maintain a tamper-proof audit trail of every agent execution—all from a single surface.

MandateOS workspace overview showing mandate count, receipts, audit events, integrity status, and setup checklist.
Workspace overviewMandates, receipts, audit events, and integrity status on one surface.
MandateOS mandate composer with Basics, Risk, Surface, and Review tabs active on the Review step.
Mandate composerThe Mandate Composer includes pre-configured presets that catch common agent risks out of the box. You can also configure your own mandates to tailor your exact need.
Mandate detail page showing the signed Mandate DSL policy block including purpose, spend cap, boundary, tools, and approval rules.
Mandate detailInspect the signed DSL, fingerprint, and policy body for any active mandate.
Workspace evidence view with recent mandates, selected mandate signature, and receipt counts.
Audit and evidenceSelected mandate evidence, receipt counts, and the wider workspace ledger together.

Where it fits

Use one approval and evidence model across the host tools your team already runs.

MandateOS sits between the agent and the actions that matter: shell execution, code changes, approvals, and higher-risk tool use.

Available today: Codex, Cursor, Claude Code, OpenClaw, and managed MCP flowsSame mandate, approval, and receipt model across local hostsPlanned: GitHub-side enforcement

Codex

Available Today

Use project-scoped Codex MCP and hook config to bring Bash-side actions under the same approval and receipt model.

Cursor

Available Today

Bring Cursor sessions under the same approval, tool-scope, and receipt model as the rest of your agent operations.

Claude Code

Available Today

Use the same operating model in Claude Code so teams do not have to reinvent policy for each shell.

OpenClaw

Available Today

Designed for OpenClaw’s flexible host surface so teams can keep explicit tool boundaries, approvals, and receipts around local agent power.

GitHub Enforcement

Planned

Repository-side actions are a natural next extension of the same policy and approval model.

Start with a concrete workflow

Install MandateOS into one repo, then watch the first approval and receipt happen.

The fastest way to evaluate MandateOS is to wire one host into one real repo, define one clear mandate, and inspect the evidence trail yourself.

Open Install Guide Inspect the Public Repo