Playwright automated web testing from English descriptions. Describe test flows in plain language, get executable test specs. Invoke with /webtest.
Installation
Details
Usage
After installing, this skill will be available to your AI coding assistant.
Verify installation:
npx agent-skills-cli listSkill Instructions
name: webapp-testing description: > Playwright automated web testing from English descriptions. Describe test flows in plain language, get executable test specs. Invoke with /webtest.
Webapp Testing
Prerequisites
Requires Playwright MCP server configured in mcp.json, or Playwright installed locally.
Commands
/webtest <description>— Generate and run a test from a plain English description/webtest record <url>— Open a browser and record user actions as a test/webtest run <file>— Run an existing test file/webtest report— Show results from the last test run
Procedure
Phase 1: Understand the Flow
Parse the user's plain English description into test steps:
Input: "Log in with test credentials, navigate to settings, change the display name, verify it saved"
Output steps:
- Navigate to login page
- Fill email and password fields
- Click login button
- Wait for dashboard to load
- Click settings link
- Clear display name field
- Type new display name
- Click save button
- Verify success message appears
- Verify display name updated
Phase 2: Generate Test
Write a Playwright test spec:
import { test, expect } from '@playwright/test';
test('update display name in settings', async ({ page }) => {
await page.goto('/login');
await page.fill('[name="email"]', process.env.TEST_EMAIL);
await page.fill('[name="password"]', process.env.TEST_PASSWORD);
await page.click('button[type="submit"]');
await page.waitForURL('/dashboard');
// ... remaining steps
});
Phase 3: Run
Execute the test using either:
- Playwright MCP server (browser_navigate, browser_click, browser_fill_form, etc.)
- Local Playwright CLI:
npx playwright test
Phase 4: Report
For each test:
- Status: PASS / FAIL
- Duration
- Screenshots on failure
- Error details with line reference
- Suggested fix if failed
Test Patterns
- Login flows
- Form submissions
- Navigation and routing
- CRUD operations
- Error state handling
- Responsive layout verification
MCMAP Usage
Primary test targets:
- Power Apps portal pages
- Copilot Studio web chat widget
- SharePoint document libraries
- Dataverse model-driven apps (via browser automation)
Safety Rules
- Never use production credentials in tests — use test/sandbox accounts only
- Never modify production data — always target sandbox environments
- Store test credentials in environment variables, never in test files
More by kevcofett
View allScans for security issues, hardcoded secrets, missing error handling, and vibe coding antipatterns. Produces PASS/FAIL verdicts with file:line references. Invoke with /review.
Unified document creation routing to docx/xlsx/pptx/pdf generation. Professional formatting, no emojis, enterprise-ready output. Invoke with /docs.
Identify code smells, complexity hotspots, and refactoring opportunities. Prioritize by impact and suggest concrete fixes. Invoke with /refactor.
Connect Claude to 500+ apps through a single MCP gateway server. Manage connections, check status, and list available integrations. Invoke with /rube.
