ark-setup-cluster
Skill439 starsChanged 7 months agoApache-2.0View on GitHub
Set up and install the Ark platform in a Kubernetes cluster. Supports default mode (existing cluster) and isolated e2e testing mode (dedicated Kind cluster). Use when the user wants to install, deploy, test, or configure Ark.
- Installs packages
In this repository29 other files it gives its agents.
- ark/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- tests/CLAUDE.mdCLAUDE.md
- a2a-protocolSkill
- Ark AnalysisSkill
- ark-architectureSkill
- ark-controller-developmentSkill
- ark-dependabot-managementSkill
- ark-sdk-developmentSkill
- ark-chainsaw-testingSkill
- Ark Dashboard and UI TestingSkill
- ark-documentationSkill
- issue-creationSkill
- ark-issuesSkill
- multi-branch-cluster-setupSkill
- openspec-apply-changeSkill
- openspec-archive-changeSkill
- openspec-bulk-archive-changeSkill
- openspec-continue-changeSkill
- openspec-exploreSkill
- openspec-ff-changeSkill
- openspec-new-changeSkill
- openspec-onboardSkill
- openspec-sync-specsSkill
- openspec-verify-changeSkill
- ark-pentest-issue-resolverSkill
- ark-researchSkill
- slack-announcementSkill
- ark-vulnerability-fixerSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 11
Isolated e2e: Create a dedicated Kind cluster for testing without affecting existing clusters
- Distinctive: line 46
Wait for user to choose, or if called by the build manager for e2e testing, default to Option B.
- Distinctive: line 89
If specific chart versions aren't published yet (e.g., on a feature branch), install components individually:
- Distinctive: line 108
Wait until all pods show Running and Ready.
- Distinctive: line 112
After testing is complete:
- Very specific: line 127
If docker info fails, Kind cannot create clusters.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Ark Setup Cluster
Set up and install the Ark platform using ark-cli. Supports two modes:
- Default: Deploy Ark to an existing cluster
- Isolated e2e: Create a dedicated Kind cluster for testing without affecting existing clusters
When to use this skill
- User wants to install or set up Ark
- Build manager needs an isolated cluster for e2e testing
- User needs to deploy Ark to their local cluster
- User needs to troubleshoot Ark installation issues
Prerequisites
- Docker - Required for Kind cluster creation
- kubectl - Kubernetes CLI
- Helm - For installing Ark components
- Node.js - For building the ark-cli tool
- Kind - For creating local clusters (
which kind)
Step 0: Check existing cluster state
CRITICAL: Always check before creating or modifying clusters.
kubectl cluster-info 2>/dev/null
kubectl get pods -n default 2>/dev/null | head -10
kind get clusters 2>/dev/nullIf a cluster is already running
Never silently destroy an existing cluster. It may have user data, demos, or in-progress work. Present options:
- Option A: Reuse existing cluster — deploy Ark into it. Fastest, but overwrites current Ark state.
- Option B: Create isolated cluster (recommended for e2e testing) —
kind create cluster --name ark-e2e-test. Fully isolated, preserves existing cluster. - Option C: Replace existing cluster — delete and recreate. Only if user explicitly confirms.
Wait for user to choose, or if called by the build manager for e2e testing, default to Option B.
Step 1: Create cluster
Default mode (existing cluster or new ark-cluster)
kind create cluster --name ark-clusterIsolated e2e testing mode
kind create cluster --name ark-e2e-test
kubectl config use-context kind-ark-e2e-testConfigure kubeconfig (if running inside Docker/DinD)
CONTROL_PLANE_IP=$(docker inspect -f '{{range.NetworkSettings.Networks}}{{.IPAddress}}{{end}}' <cluster-name>-control-plane)
kind get kubeconfig --name <cluster-name> --internal | sed "s/<cluster-name>-control-plane/$CONTROL_PLANE_IP/g" > ~/.kube/config
kubectl cluster-infoOn Mac with Docker Desktop, the default kubeconfig works without IP replacement.
Step 2: Build ark-cli from source
cd agents-at-scale-ark
npm install
cd tools/ark-cli
npm install
npm run buildStep 3: Install Ark
node tools/ark-cli/dist/index.js install --yes --wait-for-ready 5mIf specific chart versions aren't published yet (e.g., on a feature branch), install components individually:
helm install ark-controller oci://mckinsey-ark-helm.jfrog.io/ark-helm/ark-controller --version <version>
# ark-broker must precede ark-tenant: the tenant chart refuses a fresh install
# into a namespace with no broker Service.
helm install ark-broker oci://mckinsey-ark-helm.jfrog.io/ark-helm/ark-broker --version <version>
helm install ark-tenant oci://mckinsey-ark-helm.jfrog.io/ark-helm/ark-tenant --version <version>
helm install ark-api oci://mckinsey-ark-helm.jfrog.io/ark-helm/ark-api --version <version>
helm install ark-dashboard oci://mckinsey-ark-helm.jfrog.io/ark-helm/ark-dashboard --version <version>Step 4: Verify installation
kubectl get pods -n default
kubectl get services -n defaultWait until all pods show Running and Ready.
Step 5: Cleanup (e2e testing mode)
After testing is complete:
kind delete cluster --name ark-e2e-testRestore original kubectl context if needed:
kubectl config use-context <original-context>Troubleshooting
Docker not available
If docker info fails, Kind cannot create clusters. Report to user.
Check pod status
kubectl get pods -A -o wide | grep -E '(ark|cert-manager)'
kubectl describe pod <pod-name>View logs
kubectl logs deployment/ark-api
kubectl logs -n ark-system deployment/ark-controllerImportant
- DO NOT use
scripts/quickstart.sh— deprecated. Use ark-cli. - DO NOT use
npm install -g @agents-at-scale/ark— build from source to match the branch being tested. - CI uses K3s (Linux-only). Local Mac testing uses Kind. Generally compatible but minor differences possible.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

