Maturana: Hardware-isolated, zero-trust agent harness Maturana, a new open-source AI agent platform, launches with hardware-isolated microVMs for each agent, zero-trust security via encrypted secrets and egress control, and a shared knowledge graph. Built on Codex and a modular Rust core, it supports Firecracker on Linux and Hyper-V on Windows, aiming to provide a secure, lightweight alternative to existing agent harnesses. A secure agent harness that runs every agent in its own hardware-isolated microVM. Lightweight, fast, and completely yours to customise from Codex. Maturana is a secure, zero-trust, Codex-native AI agent platform. It has hardware isolation Firecracker/Hyper-V , encrypted secrets that never enter the VMs pipelock , egress network control, and a shared knowledge graph. Maturana is built on a small, modular Rust core, enccapculated in Codex skills. You use Codex to interact with Maturana, either through the CLI or the Windows / Mac application. Codex is therefore the control plane or you can use the CLI directly . The result is a secure, fast, and intuitive personal agent platform built on Codex. Agent VMs can run Claude Code, OpenCode, and Codex CLI. There is no shortage of agent harnesses. Most optimize for features, not security. The powerful ones have grown so large and so flexible that their attack surface is enormous — large enough that vendors now ship hardened shells just to make them safe to run. Others isolate agents in containers, which is the right instinct, but bind themselves tightly to a single vendor's ecosystem. I wanted something different: a harness I can actually read, that is secure because of how it is built , not because of a wall of permission checks bolted on afterward. I also just enjoy engineering with Codex. So Maturana is a lean harness on the Codex ecosystem, with hardware-level isolation for every agent — Firecracker on Linux, Hyper-V on Windows — and zero-trust wired through the whole thing. It combines the elegance of Unix, the agentic workflow of Codex, and the isolation of a hypervisor. The core is a small Rust runtime; everything else is a skill or a tool you can read, swap, or write yourself. The name nods to Humberto Maturana, who coined autopoiesis — self-producing systems — which is what these agents do as they build their own tools and memory. Maturana is not a chat UI competing with Codex, a generic multi-control-plane framework, Docker orchestration, or multi-tenant SaaS. It is a single-operator, security-first agent harness. Secure by design. Agents are isolated with real hardware virtualization — a Firecracker or Hyper-V microVM per agent — for maximum security, not just a container. Zero-trust. Pipelock keeps secrets like API keys and credit-card numbers out of an agent's reach, and an egress proxy controls exactly which systems it can talk to. Treat every agent as an adversary and lock it down. Build anything. Maturana is built on and for Codex, the premier OpenAI agent-engineering environment. Everything is a skill — from agent creation to tools — so you customise your agents with prompts and nothing else. Self-evolving. An internal WASM engine lets agents build their own tools on the fly, safely sandboxed with no ambient authority. Shared knowledge. Maturana ships with a shared knowledge graph that scales past markdown files. Agents build their own memory, so you don't have to. Lean and fast. Maturana is built in Rust with a modular core from the start. Skills are extensions to that core — you run only what you need. One line. It downloads the prebuilt maturana binary no Rust toolchain needed , verifies its SHA-256 against the published checksum, clones the repo for the skills/examples, and registers the runtime plane as a service. Linux — control plane only curl -fsSL https://www.maturana.sh/install.sh | bash Linux that will also RUN isolated agents — add the Firecracker microVM host curl -fsSL https://www.maturana.sh/install.sh | bash -s -- --firecracker Windows Hyper-V — self-elevates once, prompts for your Windows password for the no-login boot tasks irm https://www.maturana.sh/install.ps1 | iex Build from source instead with --from-source Linux / -FromSource Windows . Uninstall any time with scripts/uninstall.sh / scripts/uninstall-windows.ps1 — add --purge / -Purge to also delete your agents and secrets. Verify it worked — open a new shell, then: maturana --help resolves on PATH maturana service status the runtime plane is registered and healthy Most first-run failures are virtualization not being available: on Linux, no /dev/kvm means enabling virtualization in BIOS and use --firecracker only on a KVM-capable host ; on Windows, enable Hyper-V first Windows 11 Pro / Enterprise / Workstations only . Maturana is Codex-native — you don't hand-assemble an agent from CLI flags. You tell Codex to build one, and it runs the maturana-agent-create skill as a guided setup wizard: it interviews you the agent's name, who you are, how you'll reach it, what it can do , writes its IDENTITY.md / SOUL.md / MATURANA.md , then launches it into a Firecracker microVM and validates a live turn — driving the maturana-agent-create → -launch → -validate skills end to end. That conversation is the product. 1. Open a fresh login shell so the kvm group + ~/.local/bin PATH apply sanity: ls -l /dev/kvm is group-readable, maturana --help resolves . 2. Log in to the harness your agent will run on at least one : codex login or: claude then /login inside it 3. Hand Codex the wheel — it's oriented by AGENTS.md + the skills/ pack: cd ~/maturana && codex Then just tell it what you want: create and launch a new agent …or invoke the skill directly — type /skills , or $maturana-agent-create . Codex runs the wizard, builds the image, boots the microVM, and tells you when your agent is up and reachable the first build downloads/builds the guest image, ~5–15 min on a cold cache; later launches are seconds — Codex streams progress as it goes . Every skill in skills/ ships as a Codex skill under ~/.agents/skills . Note: run this in a plain shell , not inside a sandboxed agent — Firecracker needs /dev/kvm , which a sandbox hides. Rather drive the CLI yourself? The skill just orchestrates these steps. cd ~/maturana mkdir -p .maturana/host-auth && cp -r ~/.codex .maturana/host-auth/codex stage harness auth maturana setup firecracker-harnesses --agent-id codex-firecracker build image + boot microVM idempotent maturana service status plane already runs as a service — confirm it's healthy maturana agent run codex-firecracker --prompt "say hi" --wait talk to it See docs/linux-firecracker-harnesses.md /ajensenwaud/maturana/blob/main/docs/linux-firecracker-harnesses.md for the full Linux guide. Maturana is Codex-native on Windows too — you don't hand-assemble an agent from CLI flags. You tell Codex to build one, and it runs the maturana-agent-create skill as a guided setup wizard: it interviews you the agent's name, who you are, how you'll reach it, what it can do , writes its IDENTITY.md / SOUL.md / MATURANA.md , then launches it into a Hyper-V microVM and validates a live turn — driving the maturana-agent-create → -launch → -validate skills end to end. That conversation is the product. 1. Open a NEW PowerShell as Administrator. New, so the User-scope PATH the installer set picks up bin\maturana.exe; elevated, because launching a Hyper-V VM needs admin. sanity: maturana --help resolves. 2. Log in to the harness your agent will run on at least one : codex login or: claude then /login inside it 3. Hand Codex the wheel — it's oriented by AGENTS.md + the skills/ pack: cd $env:USERPROFILE\maturana ; codex Then just tell it what you want: create and launch a new agent …or invoke the skill directly — type /skills , or $maturana-agent-create . Codex runs the wizard, builds the image, boots the microVM, and tells you when your agent is up and reachable first build can take ~5–15 min on a cold cache; later launches are seconds . Every skill in skills/ ships as a Codex skill under ~/.agents/skills . Note: Hyper-V is Windows 11 Pro / Enterprise / Workstations only, and launching a VM needs an elevated shell. You don't manage the hypervisor yourself — the privileged hostd daemon installed as SYSTEM, port 47832 owns the fixed Hyper-V lifecycle, and install.ps1 already registered the no-login boot tasks , so the runtime plane and your VMs come back after a reboot without an interactive login. The installer also already prepared the Ubuntu VHDX image and the agent SSH key, so the first launch has nothing left to download. Rather drive the CLI yourself? The skill just orchestrates these steps. cd $env:USERPROFILE\maturana Stage harness auth where the spec reads it Codex shown; Claude Code → host-auth\claude-code . mkdir .maturana\host-auth -Force Copy-Item -Recurse $env:USERPROFILE\.codex .maturana\host-auth\codex maturana service status plane already runs as a service — confirm it's healthy maturana spec validate examples\MATURANA.codex-hyperv.md check before launch id: codex-demo maturana agent launch examples\MATURANA.codex-hyperv.md --apply hostd creates + boots the Hyper-V microVM, then provisions the guest maturana agent run codex-demo --prompt "say hi" --wait talk to it Re-launching an existing VM needs an explicit override: set $env:MATURANA HYPERV FORCE = "true" before --apply . A second bundled spec, examples\MATURANA.claude-hyperv.md id claude-demo , runs the Claude Code harness. If a running VM stops replying, maturana doctor then maturana setup windows-harnesses refreshes the guest workers and channels without rebuilding the VM. See docs/harness-operations.md /ajensenwaud/maturana/blob/main/docs/harness-operations.md for the full Windows / Hyper-V guide. A Maturana agent is one MATURANA.md spec — identity, runtime, VM, mounts, egress, memory, channels, schedules, snapshots. Codex writes it; you can read and edit it. Full field reference: docs/maturana-spec.md /ajensenwaud/maturana/blob/main/docs/maturana-spec.md . maturana spec validate examples/MATURANA.codex-firecracker.md check before launch maturana agent launch examples/MATURANA.codex-firecracker.md --apply maturana agent inspect codex-firecracker --live health, logs, status Talk to an agent - Console TUI: maturana tui agent picker or maturana agent chat