Guide · Agent skill
How to use the Azure Cloud Architect agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The azure-cloud-architect skill is a set of instructions, reference guides and Python scripts that lets an AI agent help you plan an Azure setup. It is aimed at developers and small teams starting a new application on Azure, and at engineers who want a Bicep starting point to adapt. After this guide you will know what the skill does, how to install it, how to ask for help in plain language, and what to check before anything goes near production.
Quick start
- 1Download and unzip the skill. Download the ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. Confirm that SKILL.md is inside the azure-cloud-architect folder.
- 2Confirm Python 3 is installed. The bundled scripts need Python 3 to run. Check this on your machine before using the scripts directly.
- 3Ask for an architecture. Describe your application, expected users, budget, compliance needs and region in a natural request that mentions Azure.
- 4Generate the Bicep template. Ask the agent for a Bicep template for your chosen pattern, or run bicep_generator.py with an arch type and output file.
- 5Review costs. Provide a JSON description of your current resources and ask for cost recommendations, or run cost_optimizer.py with --config and --json.
- 6Build and validate. Run az bicep build and az deployment group validate with your own Azure CLI login before any deployment.
- 7Work through the security checklist. Check identity, secrets, networking, encryption, monitoring and compliance items before using the setup in production.
What is an agent skill, and what does this one do?
An agent skill is a folder of instructions and supporting files that an AI agent can load when a task matches the skill's description. This one is named azure-cloud-architect and its description covers designing Azure infrastructure, creating Bicep or ARM templates, optimizing Azure costs, setting up Azure DevOps pipelines, and migrating to Azure.
When installed, the agent follows a workflow defined in SKILL.md: it gathers requirements such as expected users, budget, team experience and compliance needs, suggests an architecture pattern, drafts Bicep templates, reviews a JSON description of existing resources for cost-saving ideas, produces GitHub Actions or Azure DevOps pipeline snippets, and gives you a security checklist.
- Architecture pattern suggestions: App Service Web, Microservices on AKS, Serverless Event-Driven, and Data Pipeline
- Bicep templates for core web app resources, generated by bicep_generator.py
- Cost review of a resource description file, using fixed rules rather than live Azure pricing
- CI/CD snippets for GitHub Actions and Azure DevOps
- Security checklist covering identity, secrets, networking, encryption, monitoring and compliance
Install the azure-cloud-architect skill
Download the ZIP from this page. You have two options depending on how you use your agent.
For Claude Code, unzip the download into ~/.claude/skills/ to make it available to you across projects, or into a project's .claude/skills/ folder to limit it to that project. After unzipping, the azure-cloud-architect folder should contain SKILL.md, the references folder and the scripts folder.
In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project folder
- Claude apps: Settings → Capabilities → Skills
Check what you need before you start
The skill itself only needs Python 3 to run its bundled scripts. Deploying the templates is a separate step that needs more: an Azure subscription, the Azure CLI (az), and Azure credentials. For CI/CD you also need secrets in GitHub or Azure DevOps, namely the client ID, tenant ID and subscription ID.
The skill does not deploy to your account on its own. The documented workflow generates files and commands, and running a deployment needs your own credentials and judgement.
- Python 3 for scripts/architecture_designer.py, scripts/bicep_generator.py and scripts/cost_optimizer.py
- Azure subscription and Azure CLI (az) for deploying templates
- Azure credentials, plus client ID, tenant ID and subscription ID secrets for CI/CD
Trigger the skill with natural requests
You do not need special syntax. Describe the job in plain language and mention Azure. Requests that match the skill's description, such as designing infrastructure, writing a Bicep template, cutting Azure costs, or setting up a pipeline, are the ones that should load it.
The more context you give, the better the first pass. The skill asks about users, budget, compliance and region, so answering those questions up front saves a round trip.
- "Design an Azure setup for a SaaS app with about 10,000 users and a budget of around 500 USD a month."
- "Write a Bicep template for an App Service web app with Azure SQL and Key Vault."
- "Review this JSON export of my Azure resources and suggest cost savings."
- "Create a GitHub Actions workflow that deploys my Bicep file to a dev resource group."
- "Give me a security checklist before we go to production on Azure."
Run the scripts yourself
The scripts can also be run directly from a terminal. The architecture designer takes an application type, a user count and a requirements object. The examples in the source use the following form.
The cost optimizer reads a resources file and can output JSON, which is useful when you want to compare results or feed them into another tool. The source output shows a monthly cost, a list of actions with estimated savings and priorities, and a total potential saving. The numbers in that example are illustrative, not current Azure prices.
- Architecture designer: --app-type, --users and --requirements
- Bicep generator: --arch-type and --output
- Cost optimizer: --config and --json
python scripts/architecture_designer.py \
--app-type web_app \
--users 10000 \
--requirements '{"budget_monthly_usd": 500, "compliance": ["SOC2"]}'Validate and deploy the generated Bicep
Treat generated templates as a starting point. Before deploying, build and validate the template with the Azure CLI. The source workflow runs az bicep build and then az deployment group validate against a resource group you name. If a deployment fails, check the deployment's error property, review the Activity Log for RBAC or policy errors, and validate again.
Common causes the source lists are missing Contributor rights on the resource group, an unregistered resource provider, a globally unique name already taken, and exceeded quota.
- Common fix: register a provider with az provider register --namespace Microsoft.Web if it is not registered
- Quota increases are requested in the Azure Portal under Subscriptions → Usage + quotas
az bicep build --file main.bicep
az deployment group validate \
--resource-group rg-myapp-dev \
--template-file main.bicepWhat people use it for
- Planning a first architecture for a small SaaS app on App Service with Azure SQL
- Drafting a Bicep starting point for a team that will review and adapt it before deploying
- Reviewing a JSON export of existing resources for right-sizing and reserved capacity ideas
- Generating a GitHub Actions or Azure DevOps pipeline snippet that deploys Bicep
- Building a pre-production checklist for identity, secrets, networking and encryption
Frequently asked questions
Is the azure-cloud-architect skill free to use?
The skill is published under the MIT licence, as stated on this page. Azure itself is not free: running the templates needs your own Azure subscription, and the costs you see depend on what you deploy.
What do I need to install the azure-cloud-architect skill?
Download the ZIP and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it in the Claude apps under Settings → Capabilities → Skills. The bundled scripts need Python 3.
Does the skill deploy resources to my Azure account?
No. The documented workflow generates files and commands. Deploying requires the Azure CLI, your own Azure credentials, and your judgement about what to run.
Does the skill need internet access?
The skill's listing says it does not use the network. It runs scripts, which the listing says need Python 3.
Are the cost estimates accurate?
Not as live figures. The cost script uses fixed rules rather than live Azure pricing, so its numbers are example estimates. Check actual prices in Azure before making budget decisions.
Can I use this instead of learning Bicep?
It can draft Bicep to start from, but generated templates should be reviewed and validated by someone who understands them. For production infrastructure, have an expert review the output.
How is this different from Terraform or Azure's own docs?
This page does not compare it with other tools. The skill focuses on Bicep templates, architecture pattern suggestions, cost review rules and a security checklist for Azure, and it uses Bicep as the recommended infrastructure-as-code language.
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 this skill if you are starting a new application on Azure, want a reasonable first architecture, or want a Bicep draft and a checklist to review before deploying. Pick something else, or bring in an expert, if you need production-critical infrastructure signed off, or if you expect the tool to change your Azure account directly.
Download the skill