What problem does it solve? Managing multiple Talos Kubernetes clusters (dev, prod, argocd) requires juggling separate kubeconfigs, and accidentally running commands against the wrong cluster—especially prod—can cause real damage. This Skill enforces explicit kubeconfig selection and provides the cluster inventory, safety rules, and secret-fetching workflow needed to operate the clusters correctly. ## Core Features & Use Cases - Per-environment kubeconfig routing: Maps dev, prod, and argocd clusters to their kubeconfig files, contexts, and API VIPs, defaulting to dev when unspecified and never touching prod unless asked. - Doppler-based secret retrieval: Fetches KUBECONFIG and TALOSCONFIG secrets from the Doppler talos-proxmox project with correct file permissions, or runs one-off commands via doppler run. - Cluster inventory and architecture reference: Documents node IPs, VIPs, Longhorn storage nodes, GPU operator setup, and the Argo CD app-of-apps layout so queries target the right cluster and namespace. - Use Case: Ask "why is the pod crashlooping in the monitoring namespace on prod" and the Skill routes kubectl to the prod kubeconfig, inspects logs read-only, and confirms before any mutating action. ## Quick Start Ask the assistant to list all pods across namespaces on the dev cluster using the talos-dev kubeconfig.
