Pages designed to trigger specific failure analysis classifications and root causes
Real application bugs — the agent should NOT fix the test
Counter that doesn't increment when clicked. Test verifies count increases but it never does.
Settings form shows "saved" toast but values don't persist on reload.
"Sort A-Z" button sorts Z-A instead. Incorrect application logic.
Delete removes item from DOM but shows no confirmation message.
Pricing page missing the "Premium Plan" that should be present.
Delete button shows "Deleted!" but account reappears on refresh.
The app changed but the test hasn't been updated — agent should fix the test
Cookie consent modal blocks form interaction until dismissed.
"Generate Report" button renamed to "Create Report".
Table CSS classes changed from old selectors to new ones.
Product grid changed from 3 columns to 2 columns.
Settings moved from sidebar navigation to top nav tabs.
Dropdown options changed, causing downstream step failures.
Checkout flow simplified — "Confirm Address" step removed.
"Reset Data" cleanup button was removed from the admin panel.
Pages that behave differently on each load — simulates intermittent failures
Elements get random CSS class names on each load. Class-based selectors break intermittently.
Content loads after random 0-15s delay. Tests with fixed timeouts fail intermittently.
List items shuffle on each page load. Positional selectors get wrong items.
Promotional modal appears ~50% of the time, blocking form interaction.
Content is removed and re-rendered after 500ms, causing stale element references.
Test configuration problems — wrong URLs, missing credentials, expired sessions
Points to a page that doesn't exist (404). Tests starting URL misconfiguration.
Page requires token parameter. Without credentials, shows "Access Denied".
Login form with credential validation. Wrong password shows "Invalid credentials".
Page validates session tokens. Expired sessions show "Session Expired" error.
These pages are designed for testing the AI failure analysis agent. Each page triggers a specific failure classification and root cause category when used with appropriately designed test cases.