How to use the AWS CDK Python Setup agent skill

What is an agent skill, and what does this one do?

An agent skill is a folder with a SKILL.md file that gives an AI agent specific instructions for a task. The agent reads these instructions when your request matches them. The skill itself only supplies instructions; the agent runs commands on your machine only if you ask it to.

The aws-cdk-python-setup skill covers the setup path: configuring prerequisites, creating a new CDK project, installing Python dependencies, and using the core commands synth, diff, deploy and bootstrap.

• Walks through installing the AWS CDK CLI with npm and checking it with cdk --version
• Explains configuring AWS credentials with aws configure
• Creates a new project with cdk init app --language python
• Installs dependencies from requirements.txt into a virtual environment
• Covers cdk synth, cdk diff, cdk deploy and cdk bootstrap, plus cdk doctor for troubleshooting

What do you need before you start?

The skill assumes you already have a few tools and an AWS account. Check these before asking the agent to begin, because the steps assume familiarity with terminals and AWS credentials.

• Node.js 14.15.0 or later, required for the AWS CDK CLI
• Python 3.7 or later, used for writing CDK code
• AWS CLI, which manages credentials and resources
• Git, for version control and project management
• An AWS account with an access key and secret key

How do you install the skill?

Download the ZIP file for the skill from this page and unzip it. The folder should contain SKILL.md. Then choose one of the following locations.

For a personal skill available in all your projects, unzip it into ~/.claude/skills/. For a skill that only applies to one project, unzip it into that project's .claude/skills/ folder.

In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

• Personal: ~/.claude/skills/
• Single project: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

How do you trigger the skill with a natural request?

You do not need a special command. Describe what you want in plain language that matches the skill's purpose, such as setting up CDK or starting a Python CDK project. The agent will then use the skill's checklist.

Be specific about your situation. Mention whether you are starting fresh or fixing an existing environment, and which operating system you use, so the agent follows the matching steps. The source gives macOS/Linux and Windows activation commands, so naming your system helps.

• "Set up an AWS CDK project in Python for a new web app."
• "My cdk deploy is failing. Check my AWS credentials and region."
• "Walk me through installing the CDK CLI and checking the version."
• "Create a new CDK Python project and install its dependencies."

What does a typical setup look like?

The skill follows a fixed order. The agent installs the CDK CLI, configures credentials, creates the project with cdk init, activates the virtual environment, and installs dependencies. The setup commands from the source are shown below. The agent runs them only if you approve.

When the project is ready, the skill covers the workflow: run cdk synth to generate CloudFormation templates into cdk.out/, run cdk diff to preview changes, and then cdk deploy. Run cdk bootstrap once, on the first deployment to an environment.

• Your new project includes app.py, a folder of stack definitions, requirements.txt and cdk.json
• Primary dependencies are aws-cdk-lib and constructs

What are the safety notes?

This skill runs no scripts and uses no network access of its own. The risk comes from the commands it recommends. cdk deploy changes real cloud resources in your AWS account and can incur costs. The skill suggests using development accounts for testing and running cdk diff before any deployment.

Review every command the agent proposes before you approve it, especially deploy and bootstrap. Do not use this skill against a production AWS account without a review.

• Use development accounts for testing
• Run cdk diff before cdk deploy
• Confirm the target account and region before deploying
• Keep AWS credentials out of shared files and chat logs

What tips help the setup go smoothly?

Most setup problems come from credentials, region or version mismatches. If something breaks, ask the agent to check these first, and to run cdk doctor to diagnose the environment. The source also advises keeping requirements.txt pinned so builds stay consistent.

• Always activate the virtual environment before working
• Check that the default region is set correctly
• Confirm your Node.js and Python versions meet the minimums
• Pin requirements.txt for consistent builds
• Follow Pythonic naming and directory conventions