Live Test

Live Test - Web Application Testing

How It Works

Playwright CLI runs via Bash commands. Snapshots save to .playwright-cli/ as compact YAML files. Screenshots save as PNG. The agent reads only what it needs — no bloated accessibility trees in context.

IMPORTANT: Command Sequence

You MUST call playwright-cli open before any other command. All commands operate on the browser opened by open. If you get "browser is not open", run open first.

Cleanup Rule (CRITICAL)

All screenshots and snapshots are temporary artifacts for analysis only. After the test run completes (end of Phase 8), delete ALL generated files:

rm -f /tmp/playwright-test-*.png
rm -rf .playwright-cli/

Never leave test artifacts on disk. The final report captures all findings in text — the image files are only needed during analysis.

Quick Reference

# Browser lifecycle
playwright-cli open [url]             launch browser (optionally navigate to url)
playwright-cli open [url] --headed    launch in headed mode (visible window)
playwright-cli close                  close browser

# Navigation
playwright-cli goto <url>             navigate to a url
playwright-cli go-back                go back
playwright-cli go-forward             go forward
playwright-cli reload                 reload page

# Interaction
playwright-cli snapshot               get element refs as YAML
playwright-cli screenshot [ref]       capture current state (or specific element)
playwright-cli click <ref>            click element by ref
playwright-cli fill <ref> <text>      fill input field
playwright-cli type <text>            type text into focused element
playwright-cli select <ref> <val>     select dropdown option
playwright-cli hover <ref>            hover over element
playwright-cli press <key>            press keyboard key
playwright-cli check <ref>            check checkbox/radio
playwright-cli uncheck <ref>          uncheck checkbox/radio

# Viewport
playwright-cli resize <w> <h>         resize browser window

# DevTools
playwright-cli console [min-level]    read console messages (levels: log, warning, error)
playwright-cli network                list network requests

# Tabs
playwright-cli tab-list               list all tabs
playwright-cli tab-new [url]          open new tab
playwright-cli tab-select <index>     switch to tab
playwright-cli tab-close [index]      close tab

# Storage / Auth
playwright-cli state-save [filename]  save cookies + localStorage + sessionStorage
playwright-cli state-load <filename>  load saved state
playwright-cli cookie-list            list cookies
playwright-cli localstorage-list      list localStorage
playwright-cli sessionstorage-list    list sessionStorage

# Save
playwright-cli pdf                    save page as PDF

Workflow (8 Phases)

Phase 1: Target Detection

• If URL provided: use it directly
• If no URL: run git diff --name-only HEAD~3
• Filter for page files: src/app/**/page.tsx, app/**/page.tsx
• Map file paths to URLs:
  • Strip route groups like (dashboard), (auth)
  • Default base: http://localhost:3000
• If multiple pages changed, test each sequentially

Phase 2: Browser Setup & Authentication

Step 1 — Derive project name and session path:

PROJECT=$(basename $(git rev-parse --show-toplevel 2>/dev/null || pwd))
SESSION_FILE="$HOME/.claude/playwright-sessions/${PROJECT}.json"

Step 2 — Open browser with or without saved session:

If session file exists:

playwright-cli open --headed
playwright-cli state-load "$SESSION_FILE"
playwright-cli goto <url>

If no session file:

playwright-cli open <url> --headed

Step 3 — Check if auth is needed:

playwright-cli snapshot

• Look for: login forms, "Sign in" buttons, /login or /auth in URL
• If page loads normally (dashboard, app content) → auth is valid, continue

Step 4 — If login is needed (no session or expired):

• Tell the user: "This page requires login. Please log in manually in the browser window. Tell me when you're done."
• Wait for user confirmation
• After user confirms, save the session:

  mkdir -p "$HOME/.claude/playwright-sessions"
  playwright-cli state-save "$SESSION_FILE"
  

• Confirm: "Session saved for ${PROJECT}. Future tests will auto-login."

Step 5 — Take initial snapshot to understand page structure:

playwright-cli snapshot

Session file location: ~/.claude/playwright-sessions/<project-name>.json

• One file per project, reused across test runs
• Contains cookies + localStorage + sessionStorage
• Never committed to git (lives in ~/.claude/)
• To force re-login: delete the file or pass --reauth argument

Phase 3: Visual Review

playwright-cli screenshot --filename=/tmp/playwright-test-visual-review.png

• Read screenshot and analyze for:
  • Layout problems (misaligned elements, broken grids)
  • Spacing inconsistencies
  • Color/contrast issues
  • Typography problems
  • Missing states (loading, error, empty)

Phase 4: Functional Testing

playwright-cli snapshot

• Read YAML snapshot to identify interactive elements
• Test buttons, links, forms using refs:

  playwright-cli click <ref>
  playwright-cli fill <ref> "test input"
  

• Check console errors:

  playwright-cli console error
  

• Check network failures:

  playwright-cli network
  

Phase 5: Debug Issues

• Console errors → diagnose and fix:
  • React hydration → fix server/client mismatch
  • API failures → verify endpoints
  • Missing env vars → check .env.local
• Network failures:
  • 404 → check route exists
  • 500 → fix API code
  • CORS → check Next.js config

Phase 6: UI Improvements (Auto-Apply)

• Auto-fix issues found in Phase 3:
  • Spacing adjustments
  • Color consistency
  • Button styling
  • Typography fixes
• After each fix, reload and re-screenshot to verify:

  playwright-cli reload
  playwright-cli screenshot --filename=/tmp/playwright-test-after-fix.png
  

Phase 7: Responsive Testing

Test at three breakpoints:

# Mobile
playwright-cli resize 393 852
playwright-cli screenshot --filename=/tmp/playwright-test-mobile.png

# Tablet
playwright-cli resize 768 1024
playwright-cli screenshot --filename=/tmp/playwright-test-tablet.png

# Desktop
playwright-cli resize 1440 900
playwright-cli screenshot --filename=/tmp/playwright-test-desktop.png

• Check: touch targets, text readability, no horizontal scroll, layout adapts, no awkward breakpoints

Phase 8: Final Report & Cleanup

Structured summary of all findings:

• Pages tested
• Issues found & fixed
• UI improvements applied
• Console/network errors
• Responsive status (mobile/tablet/desktop pass/fail)

Cleanup (mandatory): Delete all temporary test artifacts:

rm -f /tmp/playwright-test-*.png
rm -rf .playwright-cli/

Leave browser open for manual review.

Session Management

Session files are stored at ~/.claude/playwright-sessions/<project-name>.json. They capture cookies, localStorage, and sessionStorage — covering both cookie-based auth (traditional sessions) and token-based auth (JWTs in localStorage).

Commands

# Save current browser state (after manual login)
playwright-cli state-save ~/.claude/playwright-sessions/my-project.json

# Load saved state into browser (browser must be open first)
playwright-cli state-load ~/.claude/playwright-sessions/my-project.json

# Inspect what's stored
playwright-cli cookie-list
playwright-cli localstorage-list
playwright-cli sessionstorage-list

Force Re-login

If auth is stale or user wants to switch accounts:

rm ~/.claude/playwright-sessions/<project-name>.json

Then run /test again — it will prompt for manual login.

Multiple Environments

For projects with staging vs production:

playwright-cli state-save ~/.claude/playwright-sessions/my-project-staging.json
playwright-cli state-save ~/.claude/playwright-sessions/my-project-prod.json

Fallback

If Playwright CLI is not installed, do NOT fall back to Chrome MCP tools. Instead, tell the user: "Playwright CLI is not installed. Install it with: npm install -g @playwright/cli@latest" and stop.