Playwright Enterprise Tester
Stop babysitting AI in the browser. Stop debugging flaky suites at 2am.
A battle-tested, governance-first Playwright plugin for Claude Code — discovers, authors, runs, classifies, fixes, and reports E2E tests on Laravel, Next.js, Bun, Cloudflare Workers, and any Node web app. Defensive defaults, 17 enterprise features, zero-surprise app-code edits.
Table of Contents
Why this plugin exists
Anyone who has tried to scale Playwright + AI on a real product ends up in
the same three traps:
- The AI burns tokens driving the browser live. Every session starts from
zero — no persistent strategy, no repeatable patterns, no shared memory of
what worked yesterday. A 200-spec suite becomes a token bonfire.
- The suite rots without governance. Flaky tests get retried into the
green. CSS selectors break on every Tailwind rebuild. Failures are not
classified, so nobody knows whether the test is wrong, the app is
wrong, or the environment is wrong. Worse: the AI silently patches your
production code to "make the test pass".
- Generic Playwright guides ignore your stack. Laravel CSRF tokens,
Livewire
wire:model bindings, Next.js 14 hydration races, Bun dev-server
readiness, Cloudflare wrangler dev quirks, Inertia partial reloads,
cross-browser shadow-DOM, accessibility, performance budgets — none of
that is in the docs.
This plugin is the answer we wish someone had handed us three years ago.
It encodes the operational discipline of a senior QA org as a Claude Code
plugin: a persistent skill that teaches the AI how to behave, an agent
that executes the workflow, a slash command that triggers it, and a
config file that constrains it. The result is an AI that authors and
diagnoses tests like a seasoned engineer — and never silently rewrites
your application code.
The 60-second value proposition
