Generate Cucumber BDD tests with Gherkin scenarios

Cucumber BDD Tests turns behavior requirements into Gherkin feature files and step definitions. The skill targets Cucumber, Gherkin, Feature and Scenario requests, and Given/When/Then workflows. It describes generation in Java, JavaScript, or Ruby, with concrete Java and JavaScript examples for browser-based login behavior. The skill is MIT licensed.

Feature files express successful and unsuccessful outcomes in business terms. A Background holds the shared visit to the login page; individual scenarios enter credentials and check either dashboard access or an invalid-credentials error. Scenario Outline and Examples separate the behavior from a table of user-specific inputs and expected results. This provides a pattern for varying data without rewriting the whole scenario.

Use the skill when a test needs to communicate behavior through readable business steps. It addresses Cucumber suites where feature files and executable bindings must stay aligned. Prefer meaningful business actions over one step per line of implementation, and declarative scenarios over imperative scripts. Repeated setup belongs in Background rather than copied across scenarios.

Avoid placing interface selectors or other low-level UI details in Gherkin. Keep those details inside step definitions where the implementation can change without rewriting the behavior narrative. Before applying generated changes, verify the code, dependencies, credentials, commands, and external service behavior in the local project. Examples remain starting points that need environment-specific checks.

Inputs include the feature goal, shared preconditions, actions, expected outcomes, and data variations. Outputs include a feature file, language-specific bindings, and supporting lifecycle or execution configuration. Java bindings use Cucumber annotations with JUnit assertions, page objects, and browser state. JavaScript bindings import Given, When, and Then from the Cucumber package, exercise page navigation and form actions, and assert content with Chai.

Before hooks establish the browser. After hooks attach a screenshot when a scenario fails and close the driver. Tags organize smoke, critical, fast, slow, and regression scenarios. The supplied Maven examples select a smoke subset and exclude slow scenarios when required:

mvn test -Dcucumber.filter.tags="@smoke"
mvn test -Dcucumber.filter.tags="@smoke and not @slow"

The quick reference also covers full Java and JavaScript runs, dry runs, and generating snippets for undefined steps.

Cloud execution uses TestMu AI through the LambdaTest browser endpoint. LT_USERNAME and LT_ACCESS_KEY provide credentials, while capabilities carry build name, scenario name, platform, and video settings. Both Java RemoteWebDriver and JavaScript Builder patterns are included.

The deeper playbook covers Maven and runner setup, DataTable and outlines, typed steps, dependency injection with PicoContainer and ScenarioContext, ordered hooks, parameter transformations, DocString, thread-safe parallel execution, Allure and other reporting, GitHub Actions tag matrices, debugging, and a best-practices checklist. These topics extend the feature-and-binding workflow into shared state, execution, and reporting without moving technical implementation details into business scenarios.