assure-quality

Type checking, linting, and other code quality checks are out of scope.

Choose QA methods flexibly based on project conventions, project type, and the target scope. Select one or combine multiple methods from the following to verify behavior.

QA Methods

E2E Exploratory Testing

Web applications only. Verifies happy-path E2E behavior.

Playwright CLI Usage:

npx -y --package '@playwright/cli@latest' -- playwright-cli ...

(referred to as playwright-cli below)

Target URL:

• Follow project configuration for the URL to access
• If a local dev server is needed, try accessing first (it may already be running in another terminal). Only start the server yourself if the connection fails.

Procedure:

# 1. Check CLI usage
playwright-cli --help

# 2. Launch browser
playwright-cli open

# 3. Start recording
playwright-cli video-start

# 4. Navigate to the target page
playwright-cli goto 'https://example.com'

# 5. Take a snapshot to get ref IDs (required to identify interaction targets)
playwright-cli snapshot
# → outputs a DOM tree with ref IDs (e.g. ref=e10) to .playwright-cli/*.yml

# 6. Interact using ref IDs
playwright-cli click e10          # click
playwright-cli fill e20 'hello'   # text input
playwright-cli select e30 'opt1'  # select dropdown
playwright-cli hover e40          # hover
playwright-cli press Enter        # key press

# 7. Snapshot again after each action to verify state → proceed to next action
playwright-cli snapshot

# 8. Check for console errors
playwright-cli console error

# 9. Take screenshots for evidence
playwright-cli screenshot         # full page
playwright-cli screenshot e10     # specific element

# 10. Stop recording → saves video to .playwright-cli/assets/{identifier}.webm
playwright-cli video-stop

# 11. Close browser
playwright-cli close

Core loop: goto → snapshot → (read yml to identify refs) → click/fill → snapshot → ...

Notes:

• Snapshots (yml) are output to .playwright-cli/
• Videos are saved to .playwright-cli/assets/{identifier}.webm
• Use -s=<name> to manage multiple browser sessions
• state-save auth.json / state-load auth.json to persist/restore auth state
• route '*/api/*' to mock network requests

Automated Testing

1. Identify existing tests related to the target code
2. Review test case coverage — pay special attention to error cases, boundary values, and semi-normal scenarios
3. If gaps are found, implement additional tests
4. Run all relevant tests and confirm they pass

Report

Unless a specific output destination or format is specified, output the following template as the result. If instructions specify a report file path or a different format, follow those instructions instead.