| name | test-generate |
| description | Generate project-specific Playwright test files by analyzing the codebase. Use when user wants custom tests written for their app rather than running generic test suites. |
| allowed-tools | Bash(playwright-cli:*) Bash(npx:*) Bash(mkdir:*) Read Write Glob Grep |
Generate Project-Specific Tests
Analyze the user's codebase and generate tailored Playwright test files that run natively via npx playwright test at zero token cost.
Setup
URL="${ARGUMENTS:-http://localhost:3000}"
TIMESTAMP=$(date +%Y%m%d-%H%M%S)
OUTDIR="test-results/test-generate/$TIMESTAMP"
mkdir -p "$OUTDIR"
Before running, verify prerequisites per references/prerequisites.md.
Phase 1: Assess
- Read the project structure — identify framework (Next.js, Vite, React, static HTML, etc.)
- Find entry points:
src/pages/, src/routes/, app/, index.html
- Identify forms, interactive components, API calls
- Check for existing test files in
tests/, e2e/, __tests__/
- Check for existing
playwright.config.ts — if missing, will create one
Phase 2: Plan
Present the user with a test plan:
- List discovered routes/pages
- List discovered forms and interactive elements
- Propose test file organization:
tests/e2e/
happy-path.spec.ts — core user journeys
validation.spec.ts — form validation, required fields, error states
edge-cases.spec.ts — empty inputs, special characters, boundary values
accessibility.spec.ts — keyboard nav, aria labels, screen reader basics
- Wait for user approval before generating
Phase 3: Generate
Write .spec.ts files using Playwright's test runner API:
- Use
test() and expect() from @playwright/test
- Prefer
getByRole(), getByLabel(), getByText() over CSS selectors
- Each test independent — no shared state
- Use
test.beforeEach() for navigation setup
- Descriptive test names:
test('submits contact form with valid data and shows success message')
Phase 4: Configure
If no playwright.config.ts exists, generate one:
- Auto-detect dev server command (
npm run dev, npm start, etc.)
- Configure
webServer to auto-start dev server
- Set up browser projects (chromium, firefox, webkit)
- Configure reporter (line for CI, html for local)
- Set
testDir to match generated test location
Phase 5: Verify
Run the generated tests:
npx playwright test --reporter=line 2>&1
If failures occur, enter the fix loop (references/fix-loop.md).
Phase 6: Report
Generate $OUTDIR/report.md summarizing what was generated:
- Number of test files created
- Number of test cases per file
- First-run results (pass/fail)
- Fix attempt history (if fix loop was entered)
- Suggested additional coverage areas
