GKE Production Readiness Review
Assess whether GKE clusters and workloads are ready for production.
The skill requires discovery before recommendations, and its listed commands are primarily read-only; it also asks for confirmation before state-changing implementation. However, it mandates invoking many specialized skills and encourages proactive execution without defining permission boundaries, data flows, sensitive-data handling, rollback, or audit logging, so points are deducted.
The phases are coherent and cover cluster, workload, and major production domains. However, availability of the dependent skills is not verified in this file, commands use placeholders, and there are no abnormal-input rules, diagnostic failure messages, tests, or key-path reproduction evidence; the static calibration therefore limits the score.
The skill identifies single-application, multi-application, and target-cluster scenarios, distinguishes application-level from cluster-wide focus, and states a non-fit for deep single-domain work. Inputs, output format, exclusion triggers, Chinese-language support, and mainland-China network fit remain insufficiently specified, so points are deducted.
The documentation is well organized, the repository README gives installation guidance, and license and official organizational provenance are clear. The skill itself lacks dependency installation notes, sample reports, FAQs, known limitations, version/change history, and explicit maintenance responsibility, so points are deducted.
The intended result is a cross-domain production-readiness assessment with RAG statuses and an overall score, which could reduce coordination effort. However, it depends on specialized skills whose content is not supplied, defines neither report fields nor acceptance criteria, and provides no directly usable representative output evidence; only limited credit is justified.
The discovery commands and specialized-skill list provide some auditable structure, and the README corroborates the skill path. There is no committed test suite, CI coverage, third-party execution evidence, or cross-source corroboration, so the claims remain largely document-based and the score is limited.
- The mandatory use of many specialized skills and the instruction to execute them proactively may broaden the operational scope; require explicit per-action permissions, confirmation, impact scope, and rollback before changes.
- No fixed report schema, evidence-recording method, or failure procedure is defined; manually review all findings before production decisions.
- Handling of sensitive configuration, cluster metadata, and logs is unspecified; do not expose credentials, tokens, or secret configuration in reports or tool output.
- The core workflow depends on other specialized skills and Google Cloud tools, with no stated mainland-China reachability or Chinese-language guarantee.
What does this skill do, and when should you use it?
This is a high-level productionization orchestrator for GKE. It covers scalability, security, reliability, observability, backup and disaster recovery, and cost optimization. It discovers the current cluster and workload configuration, coordinates relevant domain skills, and consolidates the findings. The result includes Red-Amber-Green statuses for each area and an overall readiness score. It is intended for comprehensive pre-launch reviews, not as a replacement for deep single-domain implementation skills.
Runs gcloud container clusters describe to inspect cluster details, Autopilot status, and release channel. Uses kubectl to inspect deployments or StatefulSets, namespace labels, service accounts, resource requests and limits, health probes, HPA, PDB, and NetworkPolicies. Depending on the environment, it coordinates skills including gke-app-onboarding, gke-workload-scaling, gke-observability, gke-reliability, gke-platform-security, gke-workload-security, gke-backup-dr, gke-service-networking, and gke-cost-optimization. It then produces domain-level RAG statuses, an overall readiness score, and a remediation plan.
- A platform engineer preparing one Kubernetes application for launch needs checks for probes, HPA, resource limits, PDB, and Workload Identity.
- A team moving an application that is not yet on GKE needs a plan for containerization, image building, and initial deployment.
- A cluster operations team needs a review of cluster autoscaling, multi-zonal setup, release channels, maintenance windows, and default network policies.
- A cloud platform group wants one pre-production review covering security, reliability, monitoring, backup, recovery, and cost.
- A technical lead needs Red-Amber-Green priorities for production remediation work.
What are this skill's strengths and limitations?
- Unifies review across major production-readiness domains and returns RAG results plus an overall score.
- Supports single-application, multi-application, and target-cluster assessments.
- Provides a concrete discovery checklist for GKE and Kubernetes configuration.
- Distributed under the Apache 2.0 license.
- It is a meta-skill and orchestrator, so it does not replace deep implementation skills for individual domains.
- It depends on gcloud, kubectl, and access to the target GKE environment.
- The source provides no test suite, platform validation results, or detailed scoring rubric.
- The README documents collection installation but not the exact standalone selection and installation flow for this skill.
How do you install this skill?
Run npx skills add google/skills, then select specific skills from the repository during the installation flow. The source does not document the exact interaction for selecting skills/cloud/gke-productionize or provide a separate installation command.
How do you use this skill?
Install the skills/cloud/gke-productionize folder in an Agent Skills-compatible client and ask something like: “Assess whether the GKE cluster named prod-cluster in us-central1 and the orders service in the payments namespace are production-ready.” Prepare the required gcloud, kubectl, and GKE access beforehand. The skill requires environment discovery and relevant domain-skill execution before implementation; state-changing implementation should be confirmed by the user first.
How does this skill compare with similar options?
The SKILL.md distinguishes this orchestrator from domain-specific skills such as gke-scaling, gke-platform-security, gke-workload-security, gke-service-networking, and gke-reliability: this skill coordinates the comprehensive assessment, while those skills handle deeper domain work.