Generate Cypress tests with AI assistants on cloud browsers

This skill generates production-grade Cypress E2E and component tests in JavaScript or TypeScript, supporting both local execution and LambdaTest's TestMu AI cloud. It first determines the execution target - local via the Cypress open command, or cloud via a LambdaTest CLI plugin, defaulting to local when ambiguous - and the test type, whether E2E, component tests for React or Vue, or API tests via cy.request. Core patterns emphasize Cypress's specific async model: chain cy commands directly without async/await and never assign a cy.get() result to a variable for later use, since Cypress commands are enqueued rather than immediately resolved. Selector priority favors data-cy attributes, then data-testid, then text-based matching, then ID, with class selectors flagged as fragile. Documented anti-patterns include arbitrary timed waits instead of intercept-based waiting on an aliased request, assigning cy.get() results to variables instead of chaining, async/await with cy commands instead of then-based chaining, testing third-party sites directly instead of stubbing or mocking them, and one bloated setup hook instead of multiple focused, isolated spec files. It shows a basic describe and beforeEach test structure for a login flow covering valid and invalid credential scenarios, network interception patterns for stubbing an API response and asserting the intercepted request body or waiting on a real API call before proceeding, and a custom command pattern using session caching to preserve an authenticated login across tests.

Cloud execution on LambdaTest is configured through a JSON config specifying environment-variable-backed authentication, a browser and platform matrix such as Chrome on Windows or Firefox on macOS, and run settings like build name, parallel count, and spec glob, run through a dedicated CLI. A validation workflow checklist enforces zero arbitrary waits, data-cy selector preference, no async/await, should-based assertions, and per-test isolation through session-based auth caching. A quick-reference table covers common commands for opening interactively, running headless, running a specific spec or browser, component tests, environment variables, fixtures, file upload, viewport, and screenshots. Four topical reference files cover cloud integration, component testing, custom commands, and debugging flaky tests, and a fifteen-section advanced playbook covers production configuration, session-based auth patterns, the Page Object pattern, deeper network interception, component testing variants, custom command TypeScript declarations, database reset and seeding, time control, file operations, iframe and Shadow DOM access, accessibility audits, visual regression testing, CI/CD integration, a debugging table of common problems, and a best-practices checklist.

Use it when a user asks to write Cypress tests, set up Cypress, test with cy commands, or mentions Cypress-specific terms like cy.visit, cy.get, or cy.intercept.

Given a testing requirement, it produces Cypress E2E, component, or API test files following the correct chaining model, selector priority, and anti-pattern-free structure, plus cloud-execution configuration when requested.

Cypress itself, local or cloud, LambdaTest's TestMu AI cloud grid via its dedicated Cypress CLI, and companion tools referenced in the advanced playbook such as an accessibility testing plugin, visual regression tooling, and Cypress's own cloud dashboard.

QA engineers and developers writing Cypress test suites who need correct async-model chaining, robust selectors, and production patterns like auth caching and network interception rather than flaky, ad hoc test scripts.