gke-app-onboarding

GKE App Onboarding is a Google Kubernetes Engine workflow skill that guides taking an application from source code to a running deployment on GKE for the first time. It solves the end-to-end onboarding problem — assessing the app, containerizing it, building and storing an image, generating Kubernetes manifests, and deploying — while steering toward secure, production-sensible defaults. It explicitly scopes itself to first-time onboarding and containerization, deferring general cluster administration and upgrades to sibling skills.

The workflow starts with an app assessment (language and framework, dependencies, configuration, statefulness, networking, and health endpoints), then covers containerization with best practices such as multi-stage builds, distroless or minimal base images, running as a non-root user, and logging to stdout/stderr, with Cloud Native Buildpacks as an alternative to writing a Dockerfile. It handles image management via Artifact Registry, including configuring Docker auth and enabling vulnerability scanning, then generates Deployment and Service manifests with resource requests and limits, liveness and readiness probes, and multiple replicas. Deployment is done through preferred MCP tools (apply_k8s_manifest, get_k8s_rollout_status, get_k8s_resource, and related) with a kubectl fallback, and it points to follow-on skills for autoscaling, observability, security hardening, and reliability.

It targets developers and platform engineers onboarding services to GKE. Use cases include containerizing an existing app, writing a first Dockerfile and deployment manifest, pushing to Artifact Registry with scanning enabled, and rolling out and verifying a workload. The commands are standard, security-conscious build and deploy operations against the user's own cloud project.