ConfigsJul 22, 2026·2 min read

AOS Community Edition — Open Agent Operating System

An open-source agent operating system that provides kernel-level primitives for running, isolating, and orchestrating AI agents. Built in Rust with a focus on security and composability.

Agent ready

Review-first install path

This asset needs a review step. The copied prompt tells the agent to dry-run, show the writes, then proceed only after confirmation.

Needs Confirmation · 64/100Policy: confirm
Agent surface
Any MCP/CLI agent
Kind
Skill
Install
Single
Trust
Trust: Established
Entrypoint
AOS CE
Review-first command
npx -y tokrepo@latest install 436b1ecd-85aa-11f1-9bc6-00163e2b0d79 --target codex

Dry-run first, confirm the writes, then run this command.

Introduction

AOS Community Edition is an open-source agent operating system built in Rust. It provides low-level primitives for running, sandboxing, and orchestrating AI agents with strong isolation guarantees, treating agents as first-class OS-level processes.

What AOS CE Does

  • Provides process-level isolation for running AI agents securely
  • Manages agent lifecycle including spawning, scheduling, and termination
  • Offers inter-agent communication via typed message passing
  • Enforces capability-based security policies per agent
  • Supports persistent agent state across restarts

Architecture Overview

AOS CE implements a microkernel-inspired design in Rust. Each agent runs in an isolated sandbox with explicitly granted capabilities (filesystem, network, IPC). The kernel manages scheduling, resource limits, and message routing between agents. A thin host ABI connects agent runtimes to the kernel services.

Self-Hosting & Configuration

  • Install via the official installer script or build from source
  • Initialize a workspace with aos init
  • Configure agent permissions in the workspace manifest
  • Set resource limits (CPU, memory, network) per agent
  • Deploy locally or on cloud infrastructure

Key Features

  • Capability-based security model for agent isolation
  • Typed inter-agent communication protocol
  • Agent lifecycle management with automatic restart
  • Resource accounting and quota enforcement
  • Extensible host ABI for custom runtime integrations

Comparison with Similar Tools

  • Docker — Container isolation; AOS provides agent-specific abstractions
  • Firecracker — MicroVM isolation; AOS is lighter with agent-native APIs
  • E2B — Cloud sandboxes; AOS runs locally with OS-level primitives
  • Modal — Serverless compute; AOS focuses on agent orchestration

FAQ

Q: What language are agents written in? A: Agents can be written in any language. The host ABI supports common runtimes including Python, Node.js, and Rust.

Q: How does isolation work? A: Each agent runs in a restricted sandbox with only the capabilities explicitly granted in the workspace manifest.

Q: Can agents communicate with each other? A: Yes, via typed message passing through the kernel's IPC system.

Q: Is it production-ready? A: AOS CE is under active development. Evaluate for your use case before deploying to production.

Sources

Discussion

Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.

Related Assets