mckinseyOfficial·12 Agent Skills Included

agents-at-scale-ark

Kubernetes-native platform for building and running agentic applications

Deploys and manages declarative agent workloads on Kubernetes with custom resources for agents, models, teams, and queries. Eliminates vendor lock-in and bespoke infrastructure by providing provider-agnostic model support, multi-agent orchestration, and built-in observability. Includes CLI tools, SDKs, testing patterns, and security workflows to ship production-ready agentic systems faster.

413|97

↗

npx skills add mckinsey/agents-at-scale-ark --all -g -y

Available:

Gives your AI agent the project structure, build commands, mandatory lint and test gates, and routing rules for delegating work to the specialized skills in this repository.

All Skills in This Repository (12)

Pure Emerald Level Indicators

📦 In Repo

mckinsey

ark-research

Research technology options and produce evidence-backed recommendations with curated sources.

ark-chainsaw-testing

Run Chainsaw end-to-end tests for Ark resources with a mock LLM.

ark-controller-development

Guide safe modifications to Ark Kubernetes operators and CRD resources.

ark-documentation

Organizes Ark documentation using the Diataxis framework and taxonomy.

Ark Analysis

Clone the Ark repository and summarize its architecture, CRDs, and controllers.

ark-pentest-issue-resolver

Map pentest findings to Ark security categories and generate remediation guidance.

Ark Setup

Automate Ark source setup and local Kind cluster configuration.

Ark Dashboard and UI Testing

Automate Ark dashboard UI testing with Playwright and generate PR-ready screenshots.

ark-sdk-development

Regenerate and debug TypeScript types across the ARK stack from CRDs and OpenAPI specs.

ark-architecture

Specify design requirements, data models, and API conventions for Ark feature architectures.

ark-vulnerability-fixer

Identify and mitigate Ark vulnerabilities through CVE research and patch planning.

ark-issues

Search, create, and update GitHub issues for the Ark project.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install Ark?▼

Run `npx skills add mckinsey/agents-at-scale-ark --all -g -y` in your terminal to install all skills in this suite globally.

What is Ark used for?▼

Ark is a Kubernetes-native framework for building and running agentic applications. You define agents, models, teams, and queries as declarative resources, and Ark handles orchestration, scaling, and observability.

Does Ark work with different model providers?▼

Yes. Ark is provider-agnostic and supports OpenAI, Anthropic, Google, Azure, and local Ollama models, so you can swap providers without changing your agent code.

Can Ark run end-to-end tests without real API keys?▼

Yes. Ark's chainsaw testing skills use a mock LLM server that returns scripted responses, making tests fast, deterministic, and free of API costs.

Do I need Kubernetes experience to use Ark?▼

Basic familiarity helps, but Ark provides a CLI, dashboard, and quickstart that deploy a local cluster in minutes using Minikube, Kind, or Docker Desktop.