create-agentic-playwright

Stop teaching your AI how to write tests. Hand it the rulebook.

Scaffold a production-grade Playwright + TypeScript test framework with a harness for all major AI coding agents baked in — one command, first test run green in minutes.

npm create agentic-playwright .

The . scaffolds into the current (empty) directory and names the project after the folder; pass a name like my-tests to create a subfolder instead.

Presets

npm create agentic-playwright . -- --demo          # zero questions: demo API, all AI rule trees, smoke run
npm create agentic-playwright . -- --bare          # your own app URLs, no demo smoke run
npm create agentic-playwright .                    # interactive — 4 questions
npm create agentic-playwright my-tests -- --demo   # same, but into a new ./my-tests subfolder

Other flags: --no-claude / --no-cursor / --no-copilot, --yes, --skip-install, --skip-browsers, --skip-git, --skip-smoke.

What you get

• Scaffold core — dependency-injected page objects, API testing with runtime-validated Zod contracts, Faker + Zod data factories, enums as the single source of truth.
• AI harness — a Constitution plus 17 detailed skills for Claude Code, with parallel rule trees for Cursor and GitHub Copilot (pruned at init if you don't use them).
• Quality gates — skill linting, cross-reference checks, version single-sourcing.
• Exploration-first UI testing — playwright-cli wired in for live-app exploration before any selector is written.

The template is bundled inside this package — no network fetch at init time, and the initializer version always matches the template version.

Full documentation: https://github.com/idavidov13/agentic-playwright

License

MIT