What problem does it solve? When a Kubernetes incident occurs, engineers need to perform corrective actions like restarting pods, scaling deployments, or rolling back releases, but doing so manually under pressure risks mistakes and outages. This Skill enforces a safe, repeatable remediation workflow where every action is previewed with a dry run and confirmed before execution. ## Core Features & Use Cases - Pod Restart: Delete a pod so its managing ReplicaSet, DaemonSet, or StatefulSet recreates it, with warnings for standalone pods that will not be recreated. - Deployment Scaling: Scale a deployment up or down with guardrails such as a 50-replica cap and a required confirmation flag when scaling to zero. - Deployment Rollback: Roll back to a previous or specific revision using revision history and kubectl rollout undo. - Use Case: A payment service is stuck in CrashLoopBackOff after a bad image push. Diagnose the issue, dry-run a rollback to see the target revision and image, get confirmation, then execute the rollback and verify recovery. ## Quick Start Ask the agent to dry-run a rollback of the payment deployment in the otel-demo namespace and then execute it after confirmation.
