Building with AI agents
How I write and test code now.
My stack for this
- 01Claude Code
Writes the code.
- 02OpenAI Codex
Clicks through the app and drives Playwright.
- 03Playwright, Vitest, pytest
Agents need tests to know when they're done.
- 04Stryker mutation testing
Proves the tests catch bugs.
- 05GitHub
Where it all lives.
- 06Self-hosted MCP servers
Scoped access to client systems.
Don't
- GitLab, Bitbucket and Atlassian
Mediocre tools that teams tolerate out of habit. Agents, CI and every integration land on GitHub first.
Instead: GitHub
The picks, in brief
What each one is, my verdict, and what it costs. Every name opens the full entry: when to use it and when not, the gotchas with their fixes, provisioning and alternatives.
My main tool. I haven't hand-written code in two years.
Pricing: Included in Claude Pro ($20 a month, $17 billed annually), Max ($100 or $200 a month for 5x or 20x Pro's usage) and Team seats ($25 standard or $125 premium a month, $20 or $100 billed annually). Usage resets in five-hour windows, with weekly limits on top. Or pay API rates with a Console key or through Bedrock.
For computer control, images, and as an end-to-end test orchestrator.
Pricing: Included in ChatGPT Plus ($20 a month), Pro (from $100), Business ($25 a user a month, $20 billed annually) and Enterprise. Five-hour and weekly limits apply, and image generation uses them 3–5x faster. Or pay API rates with an OpenAI key.
End-to-end and unit tests that agents can run and fix on their own.
Pricing: Free.
Tests for your tests. It mutates the code and checks your suite notices.
Pricing: Free.
Code, notes and docs. Everything integrates with it first.
Pricing: Free: unlimited private repos and 2,000 Actions minutes a month. Team: $4 per user a month, with 3,000 minutes and protected branches on private repos. Secret push protection on private repos is an add-on (GitHub Secret Protection, $19 per active committer a month).
Give agents safe, scoped access to client systems. I've started hosting my own.
Pricing: The protocol is free. You pay for whatever hosts the server: a Worker or a Lambda function, like any other small API.
Where they sit
The layers of the stack this job touches, with every tool named on this page.
AI
Models, embeddings and agent access.
Build and test
How the code gets written and checked.
Want this stack in your AWS account?
Cloudless Labs is a Sydney consultancy. We migrate on-premises systems to AWS, build serverless apps on this stack, and coach teams to ship with Claude Code and Codex.