This server hooks Claude directly into your Kubernetes and OpenShift clusters through 37+ specialized tools. It reads pod logs, tracks Tekton pipeline runs, correlates events with failures, and runs statistical anomaly detection on cluster behavior. You get Monte Carlo simulations for config changes before you apply them, predictive log analysis that spots problems early, and automated root cause analysis for failed pipelines. It also watches certificate expiry, queries resources by label, and does semantic search across logs. Reach for this when you're debugging production issues or want Claude to monitor cluster health without writing custom kubectl scripts. Works via stdio transport locally or HTTP streaming inside the cluster.
claude mcp add --transport stdio io.github.geored-lumino uvx lumino