Replies: 3 comments
-
|
That's a great vision! Integrating Playwright is definitely the way to handle the visual testing part of that inner loop. Since Squad supports the Model Context Protocol (MCP), you can absolutely use a Playwright MCP server to give the agent the ability to interact with and verify the web UI. You don't need to build this from scratch—there are several community-maintained Playwright MCP servers that expose tools for navigation, clicking, and taking screenshots. Once you plug one into your Squad configuration, the agent can use those tools to perform the manual testing steps you described as part of its task execution. It's a powerful pattern for autonomous development! |
Beta Was this translation helpful? Give feedback.
-
|
You can do a lot of this with Squad today. By starting with your PRD and making sure to ask the Squad to include Playwright tests and CI/CD to run those tests, you can get quite close to - if not all the way to - done. |
Beta Was this translation helpful? Give feedback.
-
|
This is exactly the kind of workflow Squad is built for, @Gareth064! 🎯 You can achieve this today: start with your PRD + skeleton repo, ask Squad to create GitHub Issues from the PRD and set up milestones, include Playwright tests in the initial request so testing is baked in, and let Squad pick up issues autonomously through the loop. The key is starting with tests as a first-class requirement. Check out the MCP ecosystem for Playwright implementations — you're very close to your dream workflow! ✨ |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I have a dream where I start with a PRD in my repo, plus the skeleton of my web app. I install squad. I ask it to analysis the PRD and start creating Milestones and issues in GitHub.
It then starts working on them straight away. Creating Branches, PRs, merging, deleting branches. Raising new GitHub issues when bugs are found to be picked up later etc...
BUT what I want in this inner loop is manual testing with Playwright. So the Squad is running the app and visually checking it, pressing button, entering data etc.
Is this something that is possible today with just Squad? I'd assume I need to install Playwright MCP or something?
Beta Was this translation helpful? Give feedback.
All reactions