Project Status

What KubeClaw is, what it is not, and what you should not do with it.

KubeClaw is a prototype

KubeClaw is a learning project. Its objective is to explore what a secure, scalable environment for running AI agents on Kubernetes actually requires – by building one end to end and finding out where the difficulties are.

It is not production ready. It is not beta. There is no supported release, no stability guarantee, and no security review.

What that means concretely

	Status
Maturity	Prototype / experiment
Suitable for production	No
Suitable for handling real secrets or customer data	No
Security reviewed or audited	No
API / variable stability	None – variables and layouts change without notice
Support	None – issues may go unanswered
Breaking changes	Expected, without a deprecation period

Why it exists

Agentic AI workloads execute arbitrary code with tool access. Running them without an isolation boundary is genuinely risky, and the interesting question is what a correct boundary looks like: which network controls actually hold, how much egress restriction is practical, what the operational cost is, and where the design breaks under load.

KubeClaw is an attempt to answer those questions by construction. It combines IPv6-only Hetzner infrastructure, OpenTofu and Ansible provisioning, kubeadm, Cilium network policies with FQDN egress filtering, and Cloudflare Tunnel access. Building the whole path surfaces problems that reading about it does not.

A secondary objective is preparation for the CKA certification, which is why the cluster uses kubeadm rather than a turnkey distribution.

What it is not

Not a product. There is no roadmap commitment and no support channel.
Not a reference architecture. Several decisions are made to be instructive rather than optimal.
Not hardened. The security model documented here describes the intent of the design. It has not been adversarially tested, and you should assume gaps exist.
Not a template to fork for production. More mature approaches are expected to follow in separate projects, informed by what this one gets wrong.

If you are evaluating this for real work

Please do not deploy KubeClaw to handle production traffic, real credentials, or third-party data. Read it, take the ideas, and build something you have reviewed yourself. The infrastructure it provisions is real and costs real money -- see [Cost Estimate](/kubeclaw/docs/reference/cost-estimate/) before running `tofu apply`.

Current state

The infrastructure layer is functional: network, firewalls, SSH key handling, cloud-init, and the admin/control/worker node roles all provision and come up. NAT64/DNS64, the Ansible playbooks, and the documentation site are in place.

The Kubernetes and OpenClaw layers are documented as guides but are not yet automated, and the Kubernetes-specific firewall rules are deliberately still excluded from the OpenTofu configuration. See the Roadmap for what comes next.

Versioning

Releases use v0.x and follow semantic versioning only loosely. While the major version is 0, any release may break any interface. See the changelog for what changed.