BrowserStack Cloud Testing skill
클라우드 기반 자동화, Percy를 활용한 시각적 회귀 테스트 및 관측성 도구를 사용하여 크로스 브라우저 및 크로스 디바이스 테스트를 수행합니다.
- GitHub 스타
- 248
- 파일
- 1
- 마지막 커밋
- 2026년 11월 3일
- 라이선스
- MIT
- 허용된 도구
- —
BrowserStack Cloud Testing skill은 무엇을 하나요?
BrowserStack Cloud Testing은(는) agent skill입니다. 클라우드 기반 자동화, Percy를 활용한 시각적 회귀 테스트 및 관측성 도구를 사용하여 크로스 브라우저 및 크로스 디바이스 테스트를 수행합니다.
skill의 SKILL.md에서 생성된 요약입니다.
BrowserStack Cloud Testing skill 정보
다양한 환경에서 웹 및 모바일 애플리케이션을 검증해야 하는 개발자는 이 SKILL을 사용하여 클라우드 기반 테스트를 관리할 수 있습니다. Automate 및 App Automate와 같은 도구를 사용하여 다양한 브라우저와 실제 디바이스에서 자동화된 테스트를 실행할 수 있습니다. Playwright, Selenium, Cypress, Appium과 같은 프레임워크와 통합하여 포괄적인 엔드 투 엔드, 모바일 및 성능 테스트가 가능합니다.
사용자는 Percy를 통해 시각적 회귀 테스트를 구현하여 다양한 뷰포트에서 발생하는 UI 불일치를 잡아낼 수 있습니다. 또한 BrowserStack Local 보안 터널을 통해 스테이징 또는 localhost 애플리케이션 테스트를 지원합니다. 디버깅을 위해 AI 기반 실패 분석 및 플래키 테스트 감지를 제공하는 관측성 기능을 제공합니다. 이 기능은 GitHub Actions와 같은 CI/CD 파이프라인 내에서 작동하도록 설계되어 테스트 워크플로우를 자동화하고 병렬 세션을 통해 실행을 최적화합니다.
agent가 수행할 수 있는 작업
- 1크로스 브라우저 Playwright 테스트 구성다양한 운영 체제에서 Chrome, Firefox, Safari를 대상으로 테스트를 실행하도록 Playwright 설정을 구성합니다.
- 2시각적 회귀 테스트 구현기존 자동화 스위트에 Percy 스냅샷을 통합하여 자동화된 시각적 비교를 수행합니다.
- 3로컬 스테이징 환경 테스트애플리케이션이 프로덕션에 배포되기 전에 localhost에서 실행 중인 애플리케이션을 테스트하기 위한 보안 터널을 구축합니다.
- 4AI를 활용한 테스트 실패 분석관측성 도구를 사용하여 플래키 테스트를 감지하고 테스트 실패에 대한 AI 기반 분석을 수행합니다.
skill의 SKILL.md를 기반으로 작성되었습니다. 모든 사용 사례는 해당 문서의 인용구와 대조 확인되었습니다.
BrowserStack Cloud Testing skill 설치 방법
각 명령은 skill 폴더를 클라이언트의 skills 디렉터리에 복사합니다. 클라이언트를 재시작하거나 새 세션을 시작하여 로드하세요.Claude Code에 BrowserStack Cloud Testing skill 설치
~/.claude/skills/BrowserStack-Cloud-Testing/SKILL.md
$ npx degit PramodDutta/qaskills/seed-skills/browserstack-testing ~/.claude/skills/BrowserStack-Cloud-TestingClaude Code (project)에 BrowserStack Cloud Testing skill 설치
.claude/skills/BrowserStack-Cloud-Testing/SKILL.md
$ npx degit PramodDutta/qaskills/seed-skills/browserstack-testing .claude/skills/BrowserStack-Cloud-TestingCodex에 BrowserStack Cloud Testing skill 설치
~/.agents/skills/BrowserStack-Cloud-Testing/SKILL.md
$ npx degit PramodDutta/qaskills/seed-skills/browserstack-testing ~/.agents/skills/BrowserStack-Cloud-TestingGemini CLI에 BrowserStack Cloud Testing skill 설치
~/.gemini/skills/BrowserStack-Cloud-Testing/SKILL.md
$ npx degit PramodDutta/qaskills/seed-skills/browserstack-testing ~/.gemini/skills/BrowserStack-Cloud-TestingClaude.ai에 BrowserStack Cloud Testing skill 설치
npx degit PramodDutta/qaskills/seed-skills/browserstack-testing BrowserStack-Cloud-Testing && zip -r BrowserStack-Cloud-Testing.zip BrowserStack-Cloud-Testing # claude.ai → Customize → Skills → + → Create skill → Upload a skill → BrowserStack-Cloud-Testing.zip
자동 검사
2026년 11월 14일에 실행됨전체 보고서 →아직 검증되지 않았습니다. 스킬은 점수 80, 심각한 발견 없음, 검사에 충분한 데이터가 필요합니다. 검사 방식 · 점수를 높이기 위해 수정할 항목
- 소스 저장소 연결 및 접근 가능pramoddutta/qaskills/seed-skills/browserstack-testing is public and reachablegithub.com/pramoddutta/qaskills/tree/HEAD/seed-skills/browse ↗통과
- 오픈 소스 라이선스 (OSI)MITspdx.org/licenses/MIT.html ↗통과
- 최근 1년 내 커밋 또는 릴리스Last commit 2026-11-03github.com/pramoddutta/qaskills/commits ↗통과
- 스크립트에 다운로드 후 실행이나 자격 증명 탈취가 없음No scripts in the skill folder통과
- SKILL.md에 숨겨진 지침 없음No override phrases, safeguard switches, credential requests or hidden charactersgithub.com/pramoddutta/qaskills/blob/HEAD/seed-skills/browse ↗통과
- 사전 허용 도구가 제한됨No pre-approved tools; the agent asks before using onegithub.com/pramoddutta/qaskills/blob/HEAD/seed-skills/browse ↗통과
- SKILL.md가 Agent Skills 형식을 따름name "BrowserStack Cloud Testing" is not 1-64 lowercase letters, digits and single hyphensgithub.com/pramoddutta/qaskills/blob/HEAD/seed-skills/browse ↗결함 발견
SKILL.md
GitHub에서 보기BrowserStack Cloud Testing Skill
You are an expert in BrowserStack cloud testing platform. When the user asks you to set up cross-browser testing, configure BrowserStack Automate with Playwright or Selenium, implement Percy visual testing, or optimize cloud test execution, follow these detailed instructions.
Core Principles
- Capability-driven browser selection -- Define browser and device capabilities precisely. Use BrowserStack's capability generators to avoid configuration errors.
- Parallel execution optimization -- Maximize parallel test sessions to reduce total execution time. Design tests for independence to enable safe parallelization.
- Local testing for pre-deployment -- Use BrowserStack Local to test staging environments and localhost applications through a secure tunnel before deploying to production.
- Visual regression with Percy -- Integrate Percy snapshots into existing Playwright or Selenium tests for automated visual comparison across browsers and viewports.
- Observability for debugging -- Enable BrowserStack Observability to get AI-powered failure analysis, flaky test detection, and performance insights without additional instrumentation.
- Network and geolocation simulation -- Use BrowserStack's network throttling and geolocation capabilities to test under realistic conditions for global users.
- Cost-efficient test distribution -- Balance test coverage across browsers and devices based on analytics data. Test the top 80% of user configurations, not every possible combination.
Project Structure
browserstack-tests/
config/
browserstack.config.ts
capabilities.ts
local-config.ts
tests/
e2e/
login.spec.ts
checkout.spec.ts
search.spec.ts
visual/
homepage-visual.spec.ts
product-page-visual.spec.ts
mobile/
mobile-navigation.spec.ts
mobile-checkout.spec.ts
helpers/
browserstack-client.ts
local-tunnel.ts
capability-builder.ts
percy/
percy-config.ts
snapshot-helpers.ts
scripts/
run-parallel.sh
check-session-status.ts
reports/
.gitkeep
Playwright + BrowserStack Configuration
// config/browserstack.config.ts
import { defineConfig, devices } from '@playwright/test';
const BS_CAPS = {
'browserstack.username': process.env.BROWSERSTACK_USERNAME || '',
'browserstack.accessKey': process.env.BROWSERSTACK_ACCESS_KEY || '',
'browserstack.local': process.env.BROWSERSTACK_LOCAL || 'false',
'browserstack.playwrightVersion': '1.latest',
'browserstack.debug': 'true',
'browserstack.networkLogs': 'true',
'browserstack.consoleLogs': 'info',
};
export default defineConfig({
testDir: './tests',
fullyParallel: true,
retries: 1,
workers: 5,
reporter: [
['html', { open: 'never' }],
['json', { outputFile: 'reports/results.json' }],
],
use: {
baseURL: process.env.BASE_URL || 'http://localhost:3000',
trace: 'on-first-retry',
screenshot: 'only-on-failure',
},
projects: [
{
name: 'chrome-latest-windows',
use: {
connectOptions: {
wsEndpoint: buildWSEndpoint({
...BS_CAPS,
browser: 'chrome',
browser_version: 'latest',
os: 'Windows',
os_version: '11',
name: 'Chrome Latest Windows',
build: getBuildName(),
}),
},
},
},
{
name: 'firefox-latest-windows',
use: {
connectOptions: {
wsEndpoint: buildWSEndpoint({
...BS_CAPS,
browser: 'playwright-firefox',
browser_version: 'latest',
os: 'Windows',
os_version: '11',
name: 'Firefox Latest Windows',
build: getBuildName(),
}),
},
},
},
{
name: 'safari-latest-macos',
use: {
connectOptions: {
wsEndpoint: buildWSEndpoint({
...BS_CAPS,
browser: 'playwright-webkit',
browser_version: 'latest',
os: 'OS X',
os_version: 'Sonoma',
name: 'Safari Latest macOS',
build: getBuildName(),
}),
},
},
},
{
name: 'iphone-15',
use: {
connectOptions: {
wsEndpoint: buildWSEndpoint({
...BS_CAPS,
browser: 'playwright-webkit',
device: 'iPhone 15',
os_version: '17',
name: 'iPhone 15',
build: getBuildName(),
}),
},
},
},
{
name: 'pixel-8',
use: {
connectOptions: {
wsEndpoint: buildWSEndpoint({
...BS_CAPS,
browser: 'chrome',
device: 'Google Pixel 8',
os_version: '14.0',
name: 'Pixel 8',
build: getBuildName(),
}),
},
},
},
],
});
function buildWSEndpoint(caps: Record<string, string>): string {
const encoded = encodeURIComponent(JSON.stringify(caps));
return `wss://cdp.browserstack.com/playwright?caps=${encoded}`;
}
function getBuildName(): string {
const timestamp = new Date().toISOString().split('T')[0];
const ci = process.env.CI ? 'CI' : 'local';
const commit = process.env.GITHUB_SHA?.substring(0, 7) || 'dev';
return `${ci}-${timestamp}-${commit}`;
}
BrowserStack Local Tunnel
// helpers/local-tunnel.ts
import * as BrowserStackLocal from 'browserstack-local';
export class LocalTunnel {
private bsLocal: any;
constructor() {
this.bsLocal = new BrowserStackLocal.Local();
}
async start(options?: {
key?: string;
localIdentifier?: string;
verbose?: boolean;
force?: boolean;
forceLocal?: boolean;
}): Promise<void> {
return new Promise((resolve, reject) => {
const config = {
key: options?.key || process.env.BROWSERSTACK_ACCESS_KEY,
localIdentifier: options?.localIdentifier || `local-${Date.now()}`,
verbose: options?.verbose ? '1' : '0',
force: options?.force ? 'true' : 'false',
forceLocal: options?.forceLocal ? 'true' : 'false',
};
this.bsLocal.start(config, (error: any) => {
if (error) {
reject(new Error(`BrowserStack Local failed to start: ${error.message}`));
} else {
console.log('BrowserStack Local tunnel started');
resolve();
}
});
});
}
async stop(): Promise<void> {
return new Promise((resolve) => {
if (this.bsLocal.isRunning()) {
this.bsLocal.stop(() => {
console.log('BrowserStack Local tunnel stopped');
resolve();
});
} else {
resolve();
}
});
}
isRunning(): boolean {
return this.bsLocal.isRunning();
}
}
Percy Visual Testing Integration
// tests/visual/homepage-visual.spec.ts
import { test, expect } from '@playwright/test';
test.describe('Homepage Visual Regression', () => {
test('homepage renders correctly on desktop', async ({ page }) => {
await page.goto('/');
await page.waitForLoadState('networkidle');
// Take Percy snapshot
await page.evaluate(() => {
// Percy snapshot via CLI or SDK
(window as any).__percy_snapshot_name = 'Homepage - Desktop';
});
// Fallback assertions
await expect(page.getByRole('heading', { level: 1 })).toBeVisible();
await expect(page.getByRole('navigation')).toBeVisible();
});
test('homepage renders correctly on mobile', async ({ page }) => {
await page.setViewportSize({ width: 375, height: 812 });
await page.goto('/');
await page.waitForLoadState('networkidle');
await expect(page.getByRole('heading', { level: 1 })).toBeVisible();
// Mobile hamburger menu should be visible
await expect(page.getByRole('button', { name: /menu/i })).toBeVisible();
});
});
BrowserStack Session Management
// helpers/browserstack-client.ts
interface SessionInfo {
name: string;
duration: number;
os: string;
os_version: string;
browser: string;
browser_version: string;
status: 'passed' | 'failed' | 'error';
reason: string;
public_url: string;
video_url: string;
logs: string;
}
export class BrowserStackClient {
private username: string;
private accessKey: string;
private baseUrl = 'https://api.browserstack.com';
constructor() {
this.username = process.env.BROWSERSTACK_USERNAME || '';
this.accessKey = process.env.BROWSERSTACK_ACCESS_KEY || '';
}
private get authHeader(): string {
return `Basic ${Buffer.from(`${this.username}:${this.accessKey}`).toString('base64')}`;
}
async updateSessionStatus(
sessionId: string,
status: 'passed' | 'failed',
reason?: string
): Promise<void> {
await fetch(`${this.baseUrl}/automate/sessions/${sessionId}.json`, {
method: 'PUT',
headers: {
'Content-Type': 'application/json',
Authorization: this.authHeader,
},
body: JSON.stringify({ status, reason }),
});
}
async getSession(sessionId: string): Promise<SessionInfo> {
const response = await fetch(
`${this.baseUrl}/automate/sessions/${sessionId}.json`,
{ headers: { Authorization: this.authHeader } }
);
const data = await response.json();
return data.automation_session;
}
async getBuilds(limit = 10): Promise<any[]> {
const response = await fetch(
`${this.baseUrl}/automate/builds.json?limit=${limit}`,
{ headers: { Authorization: this.authHeader } }
);
const data = await response.json();
return data;
}
async getBuildSessions(buildId: string): Promise<any[]> {
const response = await fetch(
`${this.baseUrl}/automate/builds/${buildId}/sessions.json`,
{ headers: { Authorization: this.authHeader } }
);
return response.json();
}
async getAccountUsage(): Promise<{ parallel_sessions_running: number; parallel_sessions_max: number }> {
const response = await fetch(`${this.baseUrl}/automate/plan.json`, {
headers: { Authorization: this.authHeader },
});
return response.json();
}
}
GitHub Actions Integration
# .github/workflows/browserstack-tests.yml
name: BrowserStack Cross-Browser Tests
on:
push:
branches: [main]
pull_request:
jobs:
browserstack:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Setup Node.js
uses: actions/setup-node@v4
with:
node-version: 20
- name: Install dependencies
run: npm ci
- name: Start BrowserStack Local
uses: browserstack/github-actions/setup-local@master
with:
local-testing: start
local-identifier: github-${{ github.run_id }}
- name: Run Playwright tests on BrowserStack
env:
BROWSERSTACK_USERNAME: ${{ secrets.BROWSERSTACK_USERNAME }}
BROWSERSTACK_ACCESS_KEY: ${{ secrets.BROWSERSTACK_ACCESS_KEY }}
BROWSERSTACK_LOCAL: 'true'
BROWSERSTACK_LOCAL_IDENTIFIER: github-${{ github.run_id }}
run: npx playwright test --config=config/browserstack.config.ts
- name: Stop BrowserStack Local
if: always()
uses: browserstack/github-actions/setup-local@master
with:
local-testing: stop
- name: Upload Reports
if: always()
uses: actions/upload-artifact@v4
with:
name: browserstack-reports
path: reports/
Capability Builder Utility
// helpers/capability-builder.ts
interface BrowserCapability {
browser: string;
browserVersion: string;
os: string;
osVersion: string;
deviceName?: string;
buildName: string;
sessionName: string;
}
export class CapabilityBuilder {
private baseCapabilities: Record<string, string> = {};
constructor() {
this.baseCapabilities = {
'browserstack.username': process.env.BROWSERSTACK_USERNAME || '',
'browserstack.accessKey': process.env.BROWSERSTACK_ACCESS_KEY || '',
'browserstack.debug': 'true',
'browserstack.networkLogs': 'true',
'browserstack.consoleLogs': 'info',
};
}
desktop(browser: string, version: string, os: string, osVersion: string): Record<string, string> {
return {
...this.baseCapabilities,
browser,
browser_version: version,
os,
os_version: osVersion,
resolution: '1920x1080',
};
}
mobile(device: string, osVersion: string, browser = 'chrome'): Record<string, string> {
return {
...this.baseCapabilities,
device,
os_version: osVersion,
browser,
realMobile: 'true',
};
}
withLocal(identifier: string): this {
this.baseCapabilities['browserstack.local'] = 'true';
this.baseCapabilities['browserstack.localIdentifier'] = identifier;
return this;
}
withBuild(name: string): this {
this.baseCapabilities['build'] = name;
return this;
}
withGeolocation(country: string): this {
this.baseCapabilities['browserstack.geoLocation'] = country;
return this;
}
withNetworkProfile(profile: 'no-throttling' | '4g' | '3g' | '2g'): this {
const profiles: Record<string, string> = {
'no-throttling': 'no-throttling',
'4g': '4g-lte-advanced-good',
'3g': '3g-umts-good',
'2g': '2g-gprs-good',
};
this.baseCapabilities['browserstack.networkProfile'] = profiles[profile];
return this;
}
}
// Usage examples
const builder = new CapabilityBuilder()
.withBuild('PR-123-smoke')
.withLocal('github-12345');
const chromeWindows = builder.desktop('chrome', 'latest', 'Windows', '11');
const safariMac = builder.desktop('safari', 'latest', 'OS X', 'Sonoma');
const iPhone15 = builder.mobile('iPhone 15', '17');
const pixel8 = builder.mobile('Google Pixel 8', '14.0');
Cross-Browser Test Strategy
Not every test needs to run on every browser. Design a tiered execution strategy that maximizes coverage while minimizing cost and execution time.
Tier 1 runs on every commit and includes the smoke test suite on Chrome latest. This catches obvious regressions quickly.
Tier 2 runs on pull request merges and includes the full regression suite on Chrome, Firefox, and Safari. This catches cross-browser issues before code reaches the main branch.
Tier 3 runs nightly and includes the complete test suite on all configured browsers and devices, plus performance benchmarks. This catches subtle issues that only appear in specific browser or device configurations.
Tier 4 runs weekly and includes accessibility testing across all browsers, visual regression on key pages, and mobile-specific interaction testing on real devices.
Configure BrowserStack Observability to track test results across tiers. This provides a dashboard showing which browsers have the most failures, which tests are most likely to fail on specific browsers, and which device configurations need additional coverage.
Best Practices
- Use BrowserStack's capability generator -- Avoid guessing browser and OS combinations. Use the official capability generator to produce correct configurations.
- Set meaningful build and session names -- Include CI build number, branch name, and timestamp in build names for easy identification in the dashboard.
- Enable BrowserStack Local for pre-production testing -- Test against staging and localhost environments using the secure tunnel before deploying to production.
- Parallelize based on your plan -- Check your BrowserStack plan's parallel session limit and configure workers accordingly to maximize throughput.
- Use Percy for visual regression -- Integrate Percy snapshots into critical page tests. Configure responsive widths to catch layout issues across breakpoints.
- Enable network and console logs -- Always enable logging in capabilities for debugging. Logs are invaluable when tests fail on specific browsers.
- Mark session status after test completion -- Use the BrowserStack API to mark sessions as passed or failed so the dashboard accurately reflects test results.
- Test on real devices for mobile -- BrowserStack real device testing is more reliable than emulators for mobile-specific bugs.
- Monitor session usage -- Track parallel session utilization to identify opportunities to optimize test distribution.
- Cache BrowserStack Local binary -- In CI, cache the BrowserStack Local binary to avoid re-downloading it on every run.
Anti-Patterns
- Testing every browser combination -- Testing on 50+ browser/OS combinations is wasteful. Analyze user analytics to identify the top 5-10 configurations that cover 90% of users.
- Running all tests on all browsers -- Run the full suite on Chrome, smoke tests on Firefox and Safari, and critical paths on mobile devices.
- Not using BrowserStack Local for staging -- Skipping local testing means bugs are only found after deployment. Always test pre-production environments.
- Hardcoding credentials in config files -- Use environment variables for BrowserStack username and access key. Never commit credentials.
- Ignoring flaky test detection -- BrowserStack Observability identifies flaky tests. Ignoring these leads to unreliable CI pipelines.
- Not setting timeouts appropriately -- Cloud execution is slower than local. Set generous timeouts for element waits and page loads.
- Running sequential tests on cloud -- Sequential execution wastes parallel session capacity. Always run tests in parallel on BrowserStack.
- Skipping session cleanup -- Orphaned sessions consume your parallel quota. Always clean up sessions in afterAll hooks.
- Not reviewing video recordings -- BrowserStack provides video recordings of every session. Review failed test videos before debugging code.
- Using stale browser versions -- Browser capabilities change with updates. Regularly update your browser version targets to match current releases.
이 스킬의 파일
1BrowserStack Cloud Testing skill 질문
Claude Code에 BrowserStack Cloud Testing skill을 어떻게 설치하나요?
npx degit PramodDutta/qaskills/seed-skills/browserstack-testing ~/.claude/skills/BrowserStack-Cloud-Testing를 실행합니다. Claude Code는 ~/.claude/skills(모든 프로젝트) 또는 .claude/skills(한 프로젝트)에서 skills를 로드하며, 설명이 작업과 일치할 때 skill을 사용합니다.
BrowserStack Cloud Testing skill을 사용할 수 있는 클라이언트는 무엇인가요?
Agent Skills 형식을 읽는 모든 클라이언트(Claude Code, Codex, Gemini CLI, claude.ai(ZIP으로 업로드))가 사용할 수 있습니다. skill 내 스크립트는 클라이언트가 코드 실행을 허용하는 경우에만 실행됩니다.
localhost 또는 스테이징 환경에서 실행 중인 애플리케이션을 어떻게 테스트할 수 있나요?
BrowserStack Local을 사용하여 보안 터널을 생성하면 프로덕션에 배포하기 전에 로컬 또는 스테이징 환경을 테스트할 수 있습니다.
BrowserStack으로 시각적 회귀 테스트를 수행할 수 있나요?
네, Playwright 또는 Selenium 테스트에 Percy 스냅샷을 통합하여 다양한 브라우저와 뷰포트에서 자동화된 시각적 비교를 수행할 수 있습니다.
BrowserStack은 실패한 테스트의 디버깅을 어떻게 도와주나요?
BrowserStack Observability는 추가적인 계측 없이도 AI 기반의 실패 분석, 플래키(flaky) 테스트 감지 및 성능 인사이트를 제공합니다.