
behave-skill
Generates Behave BDD tests for Python with Gherkin feature files and step implementations. Use when user mentions "Behave", "Python BDD", "Python Gherkin". Triggers on: "Behave", "Python BDD", "behave test", "Python feature file".
- Installs
- 214
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/lambdatest/agent-skills --skill behave-skill- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads behave-skill automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Behave BDD Skill
Core Patterns
Feature File (features/login.feature)
Feature: User Login
As a registered user
I want to log into the application
Background:
Given I am on the login page
Scenario: Successful login
When I enter "user@test.com" as email
And I enter "password123" as password
And I click login
Then I should see the dashboard
And the welcome message should say "Welcome"
Scenario: Invalid credentials
When I enter "wrong@test.com" as email
And I enter "wrong" as password
And I click login
Then I should see error "Invalid credentials"
Scenario Outline: Login with various users
When I enter "<email>" as email
And I enter "<password>" as password
And I click login
Then I should see "<result>"
Examples:
| email | password | result |
| admin@test.com | admin123 | Dashboard |
| bad@test.com | wrong | Error |
Step Definitions (features/steps/login_steps.py)
from behave import given, when, then
from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
@given('I am on the login page')
def step_on_login(context):
context.browser.get(context.base_url + '/login')
@when('I enter "{text}" as email')
def step_enter_email(context, text):
el = context.browser.find_element(By.ID, 'email')
el.clear()
el.send_keys(text)
@when('I enter "{text}" as password')
def step_enter_password(context, text):
el = context.browser.find_element(By.ID, 'password')
el.clear()
el.send_keys(text)
@when('I click login')
def step_click_login(context):
context.browser.find_element(By.CSS_SELECTOR, 'button[type="submit"]').click()
@then('I should see the dashboard')
def step_see_dashboard(context):
WebDriverWait(context.browser, 10).until(
EC.url_contains('/dashboard')
)
assert '/dashboard' in context.browser.current_url
@then('I should see error "{msg}"')
def step_see_error(context, msg):
error = WebDriverWait(context.browser, 5).until(
EC.visibility_of_element_located((By.CSS_SELECTOR, '.error'))
)
assert msg in error.text
Environment Hooks (features/environment.py)
from selenium import webdriver
def before_all(context):
context.base_url = 'http://localhost:3000'
def before_scenario(context, scenario):
context.browser = webdriver.Chrome()
context.browser.implicitly_wait(10)
def after_scenario(context, scenario):
if scenario.status == 'failed':
context.browser.save_screenshot(f'screenshots/{scenario.name}.png')
context.browser.quit()
Tags
@smoke
Feature: Login
@critical
Scenario: ...
behave --tags=@smoke
behave --tags="@smoke and not @slow"
Setup: pip install behave selenium
Run: behave or behave features/login.feature
Cloud Execution on TestMu AI
Set environment variables: LT_USERNAME, LT_ACCESS_KEY
# environment.py
from selenium import webdriver
import os
def before_scenario(context, scenario):
lt_options = {
"user": os.environ["LT_USERNAME"],
"accessKey": os.environ["LT_ACCESS_KEY"],
"build": "Behave Build",
"name": scenario.name,
"platformName": "Windows 11",
"video": True,
"console": True,
"network": True,
}
options = webdriver.ChromeOptions()
options.set_capability("LT:Options", lt_options)
context.driver = webdriver.Remote(
command_executor=f"https://{os.environ['LT_USERNAME']}:{os.environ['LT_ACCESS_KEY']}@hub.lambdatest.com/wd/hub",
options=options,
)
Report: behave --format json -o report.json
Deep Patterns
See reference/playbook.md for production-grade patterns:
| Section | What You Get |
|---|---|
| §1 Project Setup | behave.ini, project structure, dependencies |
| §2 Feature Files | Gherkin with Scenario Outline, data tables, Background |
| §3 Step Definitions | Type registration, API steps, common steps with PyHamcrest |
| §4 Environment Hooks | before_all/scenario/feature, screenshot on failure, DB isolation |
| §5 Page Objects | BasePage with waits, LoginPage, reusable components |
| §6 Fixtures & Test Data | DatabaseHelper, transaction rollback, JSON data loader |
| §7 LambdaTest Integration | Remote browser creation, cloud capabilities |
| §8 CI/CD Integration | GitHub Actions with Postgres, Selenium, Allure reports |
| §9 Debugging Table | 12 common problems with causes and fixes |
| §10 Best Practices | 14-item BDD testing checklist |
FAQ
What does the behave-skill skill do?
Generates Behave BDD tests for Python with Gherkin feature files and step implementations. Use when user mentions "Behave", "Python BDD", "Python Gherkin". Triggers on: "Behave", "Python BDD", "behave test", "Python feature file".
How do I install behave-skill?
Run npx skills add https://github.com/lambdatest/agent-skills --skill behave-skill in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/behave-skill/ (personal) or .claude/skills/behave-skill/ (project).
Which AI agents can use behave-skill?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is behave-skill safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains behave-skill?
It is published by lambdatest in lambdatest/agent-skills. It has 214 recorded installs.
More from lambdatest/agent-skills
playwright-skill
lambdatest/agent-skills
Generates production-grade Playwright automation scripts and E2E tests in TypeScript, JavaScript, Python, Java, or C. Supports local execution and TestMu AI cloud across 3000+ browser/OS combinations and real mobile devices. Use when the user asks to write Playwright tests, automate browsers, run cross-browser tests, test on real devices, debug flaky tests, mock APIs, or do visual regression. Triggers on: "Playwright", "E2E test", "browser test", "run on cloud", "cross-browser", "TestMu", "LambdaTest", "test my app", "test on mobile", "real device".
vitest-skill
lambdatest/agent-skills
Open the skill to read its SKILL.md instructions.
appium-skill
lambdatest/agent-skills
Open the skill to read its SKILL.md instructions.
pytest-skill
lambdatest/agent-skills
Open the skill to read its SKILL.md instructions.
selenium-skill
lambdatest/agent-skills
Generates production-grade Selenium WebDriver automation scripts and tests in Java, Python, JavaScript, C, Ruby, or PHP. Supports local execution and TestMu AI cloud with 3000+ browser/OS combinations. Use when the user asks to write Selenium tests, automate with WebDriver, run cross-browser tests on Selenium Grid, or mentions "Selenium", "WebDriver", "RemoteWebDriver", "ChromeDriver", "GeckoDriver". Triggers on: "Selenium", "WebDriver", "browser automation", "Selenium Grid", "cross-browser", "TestMu", "LambdaTest".
jest-skill
lambdatest/agent-skills
Open the skill to read its SKILL.md instructions.
Popular testing & qa skills
tdd
mattpocock/skills
Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.
grilling
mattpocock/skills
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
migrate-to-shoehorn
mattpocock/skills
Migrate test files from as type assertions to @total-typescript/shoehorn. Use when user mentions shoehorn, wants to replace as in tests, or needs partial test data.
run-train
lllllllama/rigorpilot-skills
Rigor Train skill for deep learning research repositories. Use when a documented or selected training command should be run conservatively for startup verification, short-run verification, full kickoff, or resume, with command, config, seed, log, checkpoint, status, and metric evidence written to standardized trainoutputs/. Do not use for environment setup, exploratory sweeps, speculative idea implementation, or end-to-end orchestration.
test-driven-development
obra/superpowers
Use when implementing any feature or bugfix, before writing implementation code
finishing-a-development-branch
obra/superpowers
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work