Claude Code Agent: Cloud Architect — AWS/GCP/Azure Design

What it is

This is a Claude Code agent skill that adds cloud architecture capabilities. It provides structured guidance for infrastructure design, cost optimization, and security best practices across AWS, GCP, and Azure.

The skill equips Claude Code to act as a cloud architect, helping developers design VPC layouts, select compute and storage services, configure IAM policies, and estimate costs. It covers multi-cloud patterns and provides infrastructure-as-code templates.

How it saves time or tokens

Cloud architecture decisions involve evaluating dozens of services, pricing tiers, and configuration options. This skill encodes best practices so Claude Code can produce well-architected infrastructure designs without requiring extensive back-and-forth prompting.

The skill's structured approach means fewer prompt iterations to arrive at a production-ready architecture, saving tokens on clarification and refinement cycles.

Additionally, the project's well-structured documentation and active community mean developers spend less time troubleshooting integration issues. When AI coding assistants generate code for this tool, they can reference established patterns from the documentation, producing correct implementations with fewer iterations and lower token costs.

How to use

1. Install the skill in your Claude Code project by adding the configuration to your CLAUDE.md or skills directory.

1. Describe your architecture requirements:

Design a serverless API with:
- REST endpoints for user management
- PostgreSQL database
- Authentication via JWT
- Auto-scaling to handle 10K requests/minute
- Monthly budget under $500

1. The agent produces an architecture diagram description, service selection rationale, IAM policies, and infrastructure-as-code templates.

1. Review and iterate on the design before implementing.

Example

# Generated AWS CDK architecture snippet
resources:
  api:
    type: AWS::ApiGatewayV2::Api
    properties:
      Name: user-api
      ProtocolType: HTTP
  function:
    type: AWS::Lambda::Function
    properties:
      Runtime: python3.12
      MemorySize: 256
      Timeout: 30
      Environment:
        Variables:
          DB_HOST: !GetAtt Database.Endpoint.Address
  database:
    type: AWS::RDS::DBInstance
    properties:
      Engine: postgres
      DBInstanceClass: db.t4g.micro

Related on TokRepo

• AI Tools for DevOps — Infrastructure and deployment tools
• AI Tools for Coding — AI-assisted development tools

Common pitfalls

• Accepting generated architectures without cost estimation. Always validate the projected monthly cost before implementing. The agent provides estimates, but verify with the cloud provider's pricing calculator.
• Over-engineering for small projects. Not every application needs multi-AZ, auto-scaling, and CDN. Start simple and add complexity as traffic demands it.
• Ignoring the shared responsibility model. The agent designs infrastructure, but security responsibilities vary between IaaS, PaaS, and serverless. Understand what you are responsible for securing.
• Failing to review community discussions and changelogs before upgrading. Breaking changes in major versions can disrupt existing workflows. Pin versions in production and test upgrades in staging first.