Skip to content
Get packs

Guide · Agent skill

How to use the GKE Storage agent skill: setup and examples

Updated · written with AI from the project's own documentation and checked against our catalogue data

The gke-storage skill is an Apache-2.0 reference guide that gives an AI coding agent working knowledge of storage on Google Kubernetes Engine (GKE). It covers block disks, Filestore shared file storage and Cloud Storage bucket mounts, plus how to write the Kubernetes manifests that request them. This guide explains what the skill does, how to install it, how to trigger it, and what to check before letting an agent touch a real cluster.

GKE Storage Configurationgoogle/skills · Apache-2.0

Quick start

  1. 1Download and unzip. Download the gke-storage ZIP from this page and unzip it so you have the skill folder containing SKILL.md.
  2. 2Place the folder. Move it into ~/.claude/skills/ for all projects, or into .claude/skills/ in one repository. Alternatively upload the ZIP under Settings → Capabilities → Skills in the Claude apps.
  3. 3Confirm your cluster access. Make sure kubectl and/or gcloud can reach your GKE cluster. Add the optional MCP server if you want the MCP tools.
  4. 4Make a natural request. Describe the storage you need, such as a database volume or shared files, and name the workload and size.
  5. 5Review the manifest. Check the StorageClass, access mode, zone behaviour and size in the generated YAML before applying it.
  6. 6Apply and verify. Apply the manifest to a non-production cluster first, then confirm the PVC is bound.

What is an agent skill and what does gke-storage do?

An agent skill is a folder containing a SKILL.md file of instructions that an AI agent loads when a task matches the skill's description. The gke-storage skill is written for storage configuration on GKE clusters, including persistent disks, file storage and cloud storage integration.

Once loaded, the agent knows the built-in StorageClasses and their disk types, how to write PersistentVolumeClaim and StorageClass manifests, and how to resize a volume. It also knows the zone and capacity constraints that can block stateful workloads.

  • Built-in StorageClasses: standard-rwo, premium-rwo, standard-rwx and premium-rwx
  • Block storage (ReadWriteOnce) on Persistent Disk and shared storage (ReadWriteMany) on Filestore
  • Volume expansion by patching the PVC when allowVolumeExpansion is true
  • Zone pinning for existing zonal volumes and guidance on automated disk type selection
  • Routing to other skills: google-cloud-storage-fuse for GCS FUSE mounts and gke-storage-troubleshooting for attach, mount and performance failures

How do you install the gke-storage skill?

Download the skill ZIP from this page and unzip it. The skill folder contains SKILL.md and references/attribution.md. Put the unzipped folder in one of the locations below, depending on whether you want it for every project or one project only.

To install it for yourself across all projects, unzip it into ~/.claude/skills/. To install it for a single project, unzip it into the project's .claude/skills/ folder. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

  • Personal (all projects): ~/.claude/skills/
  • Project only: .claude/skills/ inside your repository
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

How do you trigger the skill with natural requests?

You do not need to type the skill name. The agent decides to load the skill when your request matches its description, which centres on configuring GKE storage and creating PVCs. Name the storage type or the resource you want and the match is usually clear.

Requests that fit the skill include asking for a persistent volume for a database, asking for shared storage that several pods can write to, or asking for a PVC manifest. Requests about GCS FUSE mounts or failing volume attaches are routed to other skills, as the skill's description states.

  • "Create a 100Gi PVC for my Postgres pod on GKE using premium storage."
  • "Set up shared ReadWriteMany storage on GKE for three web pods."
  • "Grow the data volume for orders-db to 200Gi."
  • "Write a StorageClass that replicates across two zones for production."

What does a typical manifest look like?

For single-pod databases, the skill uses ReadWriteOnce on Persistent Disk with the premium-rwo StorageClass. The example below requests 100Gi. Shared access across pods uses ReadWriteMany on Filestore; the skill notes that the Filestore minimum is 1 TiB for the Basic tier.

The skill recommends volumeBindingMode: WaitForFirstConsumer so the disk is created in the same zone as the pod, and allowVolumeExpansion: true so you can resize later.

apiVersion: v1 kind: PersistentVolumeClaim metadata: name: database-pvc spec: accessModes: - ReadWriteOnce storageClassName: premium-rwo resources: requests: storage: 100Gi

How do you resize a volume with this skill?

Volume resizing works only if allowVolumeExpansion: true is set on the StorageClass. The skill's kubectl example patches the PVC's requested storage. Kubernetes then resizes the filesystem automatically.

When the agent has the optional MCP server, the skill prefers an MCP patch call over raw kubectl. Both act on a live resource, so review the target name and size before approving.

kubectl patch pvc <PVC_NAME> -p '{"spec":{"resources":{"requests":{"storage":"200Gi"}}}}'

What safety and access notes should you know?

The skill runs no scripts and makes no network calls of its own. It does, however, guide the agent to run gcloud, kubectl and MCP commands that change real resources. The skill's own content says commands such as kubectl patch act on live resources, so the agent should not be trusted to change a production cluster without review.

Every gcloud command the agent runs in the terminal is supposed to carry a usage attribution prefix, set inline per command, as described in references/attribution.md. The skill tells the agent not to use gcloud config set for this. For GCS mounts, the pod's Workload Identity service account needs storage.objectViewer on the bucket.

  • Required access: a GKE cluster with kubectl and/or gcloud access
  • Optional MCP server exposing apply_k8s_manifest, get_k8s_resource, describe_k8s_resource and get_cluster
  • Review every apply or patch before it runs against production
  • Excluded: database administration and replication design beyond volume provisioning

What people use it for

  • Creating a 100Gi ReadWriteOnce PVC on premium-rwo for a single-pod database
  • Setting up a ReadWriteMany Filestore volume on standard-rwx for several web pods sharing files
  • Writing a regional StorageClass with replication-type regional-pd for production high availability
  • Growing an existing PVC after confirming allowVolumeExpansion is enabled
  • Planning storage for a workload that must stay in one zone because an existing zonal PV pins it there

Frequently asked questions

How do I install the gke-storage agent skill?

Download the ZIP from this page and unzip it into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder for one project. In the Claude apps you can upload the ZIP under Settings → Capabilities → Skills.

What do I need to use the gke-storage skill?

You need access to a Kubernetes and GKE cluster through kubectl and/or gcloud. An optional MCP server exposing apply_k8s_manifest, get_k8s_resource, describe_k8s_resource and get_cluster is also listed. For GCS bucket mounts, Workload Identity and storage IAM roles on the bucket are required.

Is the gke-storage skill free?

The skill is licensed under Apache-2.0, as listed on this page. The source does not state any pricing for GKE or Google Cloud resources it may create, so check those costs separately.

Does the skill run scripts or use the network?

According to this page, the skill runs no scripts and uses no network. It does guide the agent to run gcloud, kubectl and MCP commands, which act on real cluster resources.

Can gke-storage fix a failing volume mount or attach error?

No. The skill routes attach, mount, disk performance and Cloud Storage FUSE OOM failures to a separate gke-storage-troubleshooting skill. Mounts for GCS buckets are handled by google-cloud-storage-fuse.

Does it replace a database administrator or handle replication?

No. The skill excludes database administration and replication strategies beyond volume provisioning. It covers creating and resizing the storage a database uses, not how the database replicates.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use gke-storage if you are an engineer or platform operator who needs correct PVC, StorageClass and Filestore defaults on GKE, and you will review every manifest before it reaches a cluster. Pick something else if you need live troubleshooting of attach or mount errors, database replication design, or a tool that changes production safely without human review.

Download the skill