Skip to content
Get packs

GKE Batch & HPC Workloads

Guides an agent through running batch, parallel and HPC jobs on Google Kubernetes Engine

Source Download skill
Repository
google/skills
Stars
21,782
Licence
Apache-2.0
Files
2 files

What it does

Once installed, the agent knows how to set up batch processing and high-performance computing jobs on Google Kubernetes Engine (GKE), a managed service for running containers. It can write Kubernetes Job and JobSet manifests, set up Kueue job queues, create compact-placement node pools for tightly coupled workloads, run MPI jobs through the MPI Operator, and suggest cost-saving options such as Spot VMs and scale-to-zero. It also reminds the agent to set resource limits and checkpointing to survive interruptions.

Great for

  • Engineers running large batch data pipelines or parallel jobs on GKE
  • Teams setting up shared job queues with Kueue for multiple users
  • Researchers or engineers running MPI-based simulations on Kubernetes

Skip it if

  • Standard web application deployments, which the skill itself says to route to the gke-app-onboarding skill instead
  • Anyone without a GKE cluster or Google Cloud account, since the examples only work against a real cluster and cost money to run
  • Production clusters changed without review, because the agent may run gcloud and kubectl commands that create or delete cloud resources

Try asking

  • “Create a Kubernetes Job that processes 100 work items with 10 running in parallel”
  • “Set up a Kueue ClusterQueue and LocalQueue for our data science team's batch jobs”
  • “Create a compact-placement node pool on GKE for an MPI simulation”
  • “Help me run a long ML training job on Spot VMs without losing progress”
Step-by-step guide: how to use GKE Batch & HPC WorkloadsInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the Apache-2.0 licence and an attribution note.

    gke-batch-hpc.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip gke-batch-hpc.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip gke-batch-hpc.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Review every manifest and gcloud command before the agent applies it, especially node pool creation and deletions, since these cost money and affect shared clusters
  • Ask the agent to set CPU and memory requests and limits and a backoffLimit on every Job, as the skill recommends
  • Test first in a non-production project or cluster, and pair this skill with the gke-compute-classes skill for Spot fallback patterns

What's inside

  • references/attribution.md
  • SKILL.md

Redistributed unmodified from google/skills under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.