azure-ai
A map of Azure's AI services — search over your own documents, speech turned into text, text turned into speech, forms and scans read automatically — pointing at the right tool or library for each job.
LABSCO SUMMARY
Summary
Microsoft's own Azure Skills Plugin installs most of the skills listed here (the rest come from smaller plugins in the same repository) alongside the Azure MCP Server's 200-plus tools and Foundry MCP in one step — it is a full Azure operating layer, not a skills library meant to be read on its own.
The skills split into two clusters. General Azure lifecycle work covers azure-prepare, azure-validate, azure-deploy, and azure-upgrade for an azd-based deployment end to end, azure-enterprise-infra-planner and azure-kubernetes for landing-zone and AKS design, and azure-diagnostics, azure-compliance, azure-resource-lookup, and azure-resource-visualizer for troubleshooting, security, inventory, and architecture diagrams once something is already running. Spend sits in a separate plugin in the same repository, azure-cost, whose four skills (cost-analysis, cost-estimation, cost-governance, cost-optimization) are listed here too. A second, Azure OpenAI and Foundry-specific cluster covers model deployment split by control level (preset, customize, capacity, deploy-model), fine-tuning (finetuning), and a single router across the rest of Foundry's agent and evaluation surface (microsoft-foundry). A handful of narrower skills sit outside both — azure-storage, azure-kusto, azure-messaging, azure-ai, azure-aigateway, entra-app-registration, entra-agent-id, and a Python-specific python-appservice-deploy.
This is for a team that has already decided to build on Azure and wants an agent calling real Azure and Foundry APIs instead of reciting documentation — most of these skills exist to route a request to the Azure MCP Server or Foundry MCP, and assume that live wiring is already in place. Without an Azure subscription, or without granting an agent az login credentials, none of them does anything on its own.
READ MORE
What it costs before any of it works.
Install requires Node.js 18+, the Azure CLI authenticated with az login, and, for the azd-based deployment skills, the Azure Developer CLI authenticated separately. Most of the skills listed here need an Azure account and credentials before they're useful, several more talk to Azure or Foundry MCP directly, and only a few run on guidance alone with nothing live behind them.
The skills and the MCP tools are written to be read together, not separately.
azure-compute's own description tells the agent to prefer itself "over mcp__azure__get_azure_bestpractices for VM create intents" and use its own sub-commands instead — several skills route around a same-named MCP tool this way, so installing only the MCP server, or only the skills, changes what the agent actually does compared to installing both as the plugin intends.
WHAT'S INSIDE
Skills
43 of 43
SETUP
Showing all 43 skills
AI models and agents5 skills
azure-aigateway
Puts a gate in front of your AI models, tools and agents so that everything passing through can be metered, capped and filtered instead of going straight to the model.
entra-agent-id
Gives every running AI agent its own login at the company, so each one gets in under its own name, reaches only what it is allowed to, and leaves its own trail.
foundry-iq
Turns your documents into a knowledge base that an AI agent on Microsoft Foundry can search and answer from, and looks after it from setup to cleanup.
microsoft-foundry
Everything you do with an AI agent on Microsoft's Foundry platform after the idea: build it, put it live, ask it things, score its answers, and work out why it went wrong.
Azure Local2 skills
azure-local
Azure Local puts Microsoft's cloud services on machines in your own building; this covers looking after it from planning and installing the cluster to updating it, running workloads on it, and fixing it when something fails.
azure-local-multi-rack
The big-hardware edition of Azure Local — Microsoft's cloud running on four or more pre-built racks in your own data centre — from planning the install to running workloads and watching its health. It is still in preview.
Cost4 skills
cost-analysis
Answers where your Azure money went, keeping the measured figures apart from any guess about why.
cost-estimation
Two kinds of Azure price question — where your bill is heading, and what something you have not built yet would cost — answered with every figure labelled as a forecast or a quote.
cost-governance
Keeping Azure spending in check with budgets, alerts, ownership labels and policies — remembering that a budget warns you but does not stop the spending.
cost-optimization
Hunts through your Azure account for spending that buys nothing and recommends what to change, without deleting, resizing or buying anything itself.
Getting an app onto Azure11 skills
azure-app-onboard
For when you know what you want to put on Microsoft's Azure cloud but not which Azure services to use: it looks at your idea or your code, proposes a setup with a price attached, writes the files, and deploys only after you say yes.
azure-app-onboard-prereq
Reads through your project before any move to Azure and marks each part of it pass, warning or fail on three questions: will it build, is it finished, and can Azure actually run it.
azure-cloud-migrate
Moves an app off Amazon, Google, Heroku or a Kubernetes cluster and onto Azure — first a report on what the move involves, then the code and configuration actually converted.
azure-compute
Everything about renting plain virtual machines on Azure — which size to pick, what it will cost, how to create one, and how to reserve capacity so it is there when you need it.
azure-deploy
Takes an app whose deployment has already been planned and checked, actually pushes it up to Azure, and tells you the web addresses it is now live at.
azure-enterprise-infra-planner
Turns a written description of what a company needs to run into a full network and security design for Azure, then into infrastructure code ready to deploy; nothing is created until you approve the plan.
azure-prepare
Turns an app's code into something Microsoft's Azure cloud can host — deciding which Azure services it needs and writing the files that set them up.
azure-storage
Sorts out where files, messages and tables should live in Azure, and how much you pay depending on how fast you need them back.
azure-validate
The check that stands between an app and the moment it goes live on Azure — is the setup complete, is the infrastructure described correctly, and is it allowed to do what it will try to do.
entra-app-registration
Sets your app up with Microsoft so people can sign in with their work or school account, and so the app can only reach the data you said it could.
python-appservice-deploy
Pushes a Python web app up to Azure so it is live on the internet, building whatever hosting pieces it needs along the way if they are not already there.
Kubernetes clusters2 skills
airunway-aks-setup
Takes a Kubernetes cluster you already have on Azure and walks it step by step — checking the hardware, installing the pieces, picking a serving engine — to the point where it can run an AI model.
azure-kubernetes
Decides how an Azure Kubernetes cluster should be set up before anyone creates it — the networking, the security, the machine pools, the monitoring and the cost controls, settled and written down first.
Kusto queries and graphs3 skills
azure-kusto-graph
Turns rows of records in Azure Data Explorer into a network of connections — which user logged into which machine, from which address — and then follows the paths through it.
azure-kusto-irql
Security investigations in Azure Data Explorer snapped together from ready-made pieces — pull the logins, emails or network traffic, attach who each person is, then filter and count — without having to know how the underlying tables are laid out.
azure-kusto-irql-graph
Takes the results of a query you already have in Azure Data Explorer and draws them as a picture of connected addresses, people, machines and messages, each kind with its own icon.
Once it is live10 skills
appinsights-instrumentation
Reading material on how to make an ASP.NET Core or Node.js app running on Azure report its own errors, timings and traffic, so you can watch it from a dashboard — the explanation, not the change itself.
azure-compliance
Audits an Azure account for security and compliance problems — misconfigured resources, weak settings, stored credentials quietly about to expire — and reports back what it found.
azure-diagnostics
Works out why something you are running on Azure has stopped working — a web app, a function, a container, a cluster, a virtual machine, a message queue — and says what to do about it.
azure-kusto
Azure Data Explorer stores mountains of log and sensor readings; this puts the question to it — how many, how often, what changed over the past day — and brings the answer back as a table or a chart.
azure-messaging
Fixes the errors that come up when an app talks to Azure's two messaging services, Event Hubs and Service Bus — connections dropping, message locks expiring, a processor that quietly stopped.
azure-quotas
Microsoft caps how much of each Azure service one account may run; this finds where your caps sit today and gets them raised when they are in the way.
azure-reliability
Checks whether an app hosted on Azure would keep running through a datacenter outage, shows where it is exposed, and fixes those spots once you approve.
azure-resource-lookup
A large Azure account quietly accumulates hundreds of things — machines, websites, disks, databases — and this answers what is actually in there, narrowed to one type, one label, or whatever is sitting unused.
azure-resource-visualizer
Draws a map of everything sitting in one corner of your Azure account and how the pieces connect, saved as a document you can hand to someone else.
azure-upgrade
Moves an app already running on Azure onto a newer hosting plan or service and checks it still works afterwards; it can also update old Microsoft code libraries inside Java projects.
Ungrouped6 skills
azure-kubernetes-automatic-readiness
Azure has a stricter, hands-off way of running Kubernetes clusters; this checks what you already run against its rules and lists everything that would block the move, with fixes where one can be written for you.
capacity
Before you can run an AI model on Azure, somewhere has to have room for it — this checks every region and project you can reach and ranks the ones that do.
customize
Sets up an AI model on Azure as a guided conversation, so every choice along the way is yours instead of a default picked for you.
deploy-model
One place to ask for an AI model on Azure — you describe what you want in ordinary words, and it picks the quick route, the full-control route, or a search for somewhere with room.
finetuning
Takes a general-purpose AI model and trains it on your own examples until it does your particular job better than the stock one.
preset
Getting an Azure OpenAI model running when capacity is scarce: your own region is tried first, and if it is full, the regions that still have room are found and offered.
No skills in this package match the current search or filters.
PLUGINS4 items
PLUGIN
azure
Microsoft Azure MCP and Skills integration for cloud resource management, deployments, and Azure services. Manage your Azure infrastructure, monitor applications, and deploy resources directly from your development environment.
PLUGIN
azure-kusto-graph-skills
Kusto graph analysis, IRQL security hunting pipelines, and graph visualization skills for Azure Data Explorer.
PLUGIN
azure-local-skills
Azure Local skills covering standard deployments (Azure Stack HCI, 1-16 node hyperconverged, rack-aware clusters) and multi-rack (rack scale) deployments built on Network Fabric Controller, Cluster Manager, and managed network fabric.
PLUGIN
foundry-iq-skills
Create and retrieve from a Foundry IQ knowledge base using File, Azure Blob, or ADLS Gen2, then connect an existing Prompt or Hosted Agent.
MCP SERVERS1 item
MCP SERVER
azure stdio
All Azure MCP tools in a single server. The Azure MCP Server implements the MCP specification to create a seamless connection between AI agents and Azure services.
Node.js 22 or newerINSTALL
Install
SINGLE SKILL
Install a skill by name
Run this outside your AI tool, using Node.js 22.20.0 or newer. Skills CLI 1.7.2 skips skill selection; prompts cancel installation without an interactive terminal. If choosing Project when asked about installation scope, open your project's folder before running this.
Install checklist
- Check selected tools if a picker appears; Claude Code may be unchecked under
Additional agents. ↑/↓ moves, Space toggles selections, Enter confirms. - Without a tool picker, check
Installing to:for your intended tool. Stop if it names a different tool. - If asked, choose
Symlink (Recommended)for shared files orCopy to all agentsfor separate copies; review the summary before confirming installation. Install the find-skills skill?defaults toYes, adding another skill for all projects. ChooseNoif you don't want that extra installation.
Find the skill name in the Skills tab. Open that skill's panel before copying a command: some skills need files or other skills from this package, and the panel shows the install route to use.
npx skills add microsoft/azure-skills --skill <YOUR_SKILL_NAME> --full-depthAPM
apm install microsoft/azure-skillsRun it in a terminal from your project folder. Microsoft says this one command installs the Azure Skills plugin across GitHub Copilot, Claude Code, Cursor, OpenCode, Codex and Gemini, from the package definition in the repository.
GitHub Copilot CLI
Add Microsoft's marketplace, then install the plugin
/plugin marketplace add microsoft/azure-skills/plugin install azure@azure-skillsType both lines in an interactive GitHub Copilot CLI session, not at your shell prompt. The first line is needed only the first time; the second installs the azure plugin, which brings the Azure skills and the Azure MCP Server. Sign in with az login, the Azure CLI, before asking it to work on live Azure resources.
Visual Studio Code
code --install-extension ms-azuretools.vscode-azure-mcp-serverRun it in a terminal; on a Mac, first add the code command from VS Code's Command Palette (Shell Command: Install 'code' command in PATH). It installs the Azure MCP extension, which also installs a companion extension that brings the Azure skills into GitHub Copilot; that skills step needs Git on your machine. Restart VS Code if Copilot does not find the skills, and run az login before asking about live Azure resources.
Claude Code
/plugin install azure@claude-plugins-officialType it inside a Claude Code session, not at your shell prompt. It opens the azure plugin's details from claude-plugins-official, Anthropic's own marketplace, which Claude Code adds the first time you start it in a terminal.
Gemini CLI
gemini extensions install https://github.com/microsoft/azure-skillsRun it in a terminal, not inside a Gemini CLI session. It installs Azure Skills as a Gemini CLI extension straight from the repository, and the extension starts the Azure MCP Server through npx, which needs Node.js 18 or later. Run az login before prompts that touch your Azure resources, and restart Gemini CLI if it is already open.
Codex
Add the marketplace, then install from /plugins
codex plugin marketplace add microsoft/azure-skillsRun the line in a terminal. It adds Microsoft's Azure Skills marketplace to Codex, is needed only the first time, and does not install the plugin by itself.
IntelliJ IDEA
Install the skills for Copilot in JetBrains
npx skills add https://github.com/microsoft/azure-skills/tree/main/.github/plugins/azure-skills/skills -a github-copilot -g -yRun it in a terminal. It installs the Azure skills into ~/.copilot/skills, the personal skills folder GitHub Copilot reads across all your projects; GitHub lists agent mode in JetBrains IDEs among the places skills work.