Skip to content
Browse Plugins

Starting questions

Deep Dives

AI: Claude and Codex patterns

ConfigHub: Deploy and manage

Plugins

Every public cub plugin, marked by its state. Released plugins have an install command; source builds link to their instructions.

New to cub? Install the cub CLI first. Public catalog packages pull and render anonymously, and you sign in only once a command saves or changes ConfigHub data.

Every plugin

Prefer a visual starting point? Explore your cluster with Scout or browse saved configuration with ConfigHub UIs.

Already running Argo CD or Flux? Start with its assessment Guide below. Bring your controller version, source revision, release and namespace, Secret ownership, and storage constraints. The first plan changes no cluster; review it before generating or running a handover.

PluginWhat it doesStateInstall
ConfigHub Workshopcub config, cub app, cub stack, cub fleet. Checks, composes and publishes configuration from the Catalog.Released, v0.6.61cub plugin install confighub/cub-workshop
Kubaracub kubara. Brings the add-ons Kubara manages for your clusters into ConfigHub. Plans and renders them offline, then hands the hub to approved releases, checks each cluster and hands back to Git. Workshop platform Guide. Guide: Run a Kubara platform through ConfigHub.Released, v0.5.2cub plugin install confighub/kubara-confighub

Tested on kind only. Handover needs cub and ConfigHub v0.9.0 or newer; Argo CD pulls with a read-only service account and argobot runs as a second one. Issuing their credentials takes Impersonate on each service account.

Sveltoscub sveltos. Brings a Sveltos fleet into ConfigHub with management-cluster delivery, health reporting and target facts. Deep Dive: AI fleet repair and prevention. Guide: Onboard a Sveltos fleet.Released, v0.15.0cub plugin install confighub/sveltos-confighub

Sveltos pulls with a read-only service account's token, which needs cub v0.8.13 or newer. Handover rollback and chart-rendered Secret handling remain open upstream work; review those limits before importing a fleet.

EKS inferencecub eksinf. Runs the eks-inference stack on EKS.Released, v0.13.0cub plugin install confighub/eks-inference
Installercub installer. Renders and installs configuration-as-data packages.Released, v0.1.0cub plugin install confighub/installer
Helmcub helm. Installs Helm charts as ConfigHub components.Released, v0.1.2cub plugin install confighub/cub-helm
Scoutcub scout. Explores and maps GitOps in your clusters.Released, v2.14.2cub plugin install confighub/cub-scout
Scancub scan, cub check. Checks Kubernetes and GitOps configuration for known risks.Released, cub-scan-v0.7.3cub plugin install confighub/homebrew-tap@cub-scan-v0.7.3 --name scan
ConfigHub Servercub server. Installs a self-hosted ConfigHub server.Released, v0.2.10cub plugin install confighub/cub-server
Democub demo. Seeds a ConfigHub organisation with a demo dataset.Released, v0.2.4cub plugin install confighub/cub-demo
Argo CDcub argo. Onboards an Argo CD estate into ConfigHub and hands it over without recreating anything; each generated Application becomes a Unit reading its own Space, and argobot reports live status back. Guide: Bring your Argo CD apps into ConfigHub.Released, cub-argo-v0.4.1cub plugin install confighub/examples@cub-argo-v0.4.1 --name argo

Start with plan, which needs no account and changes no cluster. Argo CD pulls as a service account, not a worker. Live status is recorded on each release by argobot (argobot.sh installs v0.2.3, on a Channel), or by cub argo status --watch --hard-refresh where argobot is not running; both make each release land at once. Needs cub v0.8.13 or newer for service-account tokens; live status needs ConfigHub v0.8.2 or newer.

Fluxcub flux. Onboards a Flux fleet into ConfigHub with one root per cluster reading its layers Space, records each layer's live status on the release it applied, from a reporter that runs on the cluster, and proposes each cluster that joins. Guide: Bring your Flux fleet into ConfigHub.Released, cub-flux-v0.5.1cub plugin install confighub/examples@cub-flux-v0.5.1 --name flux

Start with plan, which changes no cluster. Healthy is reported only for layers that check their workloads with wait or healthChecks; review bootstrap ownership before handover. Flux pulls as a service account, not a worker. fluxbot.sh runs the live-status reporter on the cluster, as one pod signed in with that account's key; cub flux status --watch is the same reporter run by hand. The Healthy gate reads what they record. Needs cub v0.8.13 or newer for service-account tokens; live status needs ConfigHub v0.8.2 or newer.