SEC.01
The playbook
The run starts read-only: the recon probes decide what happens next. The steps that follow are adapted to the distribution and state found — a real machine is rarely clean.
- 01reconk3s already installed? (then check the node and skip)
- 02reconanother Kubernetes already running? (then stop)
- 03reconCPU and RAM (k3s server: 2 cores and 2 GB minimum)
- 04reconports 6443/10250 free? 80/443 already taken (then disable Traefik and ServiceLB)?
- 05reconactive firewall (ufw or firewalld) to configure for the cluster networks
- 06reconoverlap with k3s CIDRs (10.42.0.0/16 pods, 10.43.0.0/16 services)
- 07reconcurrent default route (must be identical after the install)
- 08actDownload the official installerverify
- 09actAllow the cluster networks (only if ufw or firewalld is active)verify
- 10actInstall k3sdestructiveverify
- 11actThe node is Readyverify
- 12actProtect the kubeconfigverify
SEC.02
What it verifies
Each step ships its own check. The run only moves on when the check passes, so what the recipe reports as done is what is actually on the server.
- 01verifyDownload the official installerVerified by
head -1 /tmp/k3s-install.sh - 02verifyAllow the cluster networks (only if ufw or firewalld is active)Verified by
ufw status 2>/dev/null | grep -c '10.4[23].0.0/16' || firewall-cmd --zone=trusted --list-sources - 03verifyInstall k3sVerified by
systemctl is-active k3s && ip -4 route show default - 04verifyThe node is ReadyVerified by
k3s kubectl get nodes --no-headers | grep -c " Ready" - 05verifyProtect the kubeconfigVerified by
stat -c %a /etc/rancher/k3s/k3s.yaml
SEC.03
Guardrails
Non-negotiable instructions the run must respect, whatever it finds on the server.
Do not install it if another Kubernetes is already running. Less than 2 GB of RAM: warn and ask for confirmation. Never disable ufw/firewalld: only add the cluster CIDRs. Never expose 6443, 10250 or 8472/udp publicly. If the 10.42/10.43 CIDRs overlap an existing route, stop and ask for other CIDRs. After the install, check the default route has not changed. Never read the contents of k3s.yaml or of the token in /var/lib/rancher/k3s/server/token.
SEC.04
How a recipe runs in Servor
Servor is zero-knowledge: the platform cannot execute anything on its own. When you launch this recipe, a run reads the reconnaissance, plans the steps and asks for your approval. Each approved command is signed by your browser — with a key the server never sees — then relayed verbatim to the agent, which verifies the signature locally before executing.
search by intent — type “firewall”, get ufw
- 01Read-only reconnaissance before any change
- 02Every command signed by your browser, approved by you
- 03Each step verified before the run moves on, every run kept in the history
SEC.05
Deploy on Servor button
Maintain k3s (lightweight Kubernetes) or a project that needs it? Put this button in your README or on your site: one click and your users install it on their own server.
SEC.06
More containers recipes
- 01Docker EngineDocker Engine + Compose from the official repository, service active, hello-world verified.
- 02Docker ComposeDocker Compose plugin (docker compose), installed and verified.
- 03PodmanPodman (daemonless containers, rootless capable), verified.
- 04PortainerPortainer CE (LTS) Docker management web UI, admin preconfigured.
- 05WatchtowerAutomatic, scheduled updates for Docker containers.
Verified recipe
Run this recipe on your server
Connect a server, launch the recipe, and approve each step as Servor executes and verifies it. Free plan, no card required.