vc-mcp-testing-module

Agentic QA system for the Virto Commerce B2B e-commerce platform.

This is NOT a traditional test automation codebase with .spec.js files. Tests are executed through natural language prompts via MCP (Model Context Protocol) servers — LLM-powered browser automation with AI agents that navigate, test, and report.

Quick Start

For experienced developers who want to get running fast:

git clone https://github.com/VirtoCommerce/vc-mcp-testing-module.git
cd vc-mcp-testing-module
npm install
npx playwright install chromium firefox
# Create .env.local (secrets) → npm run env:check
# Create .mcp.json (see Step 4 below) → restart IDE
# Open Claude Code → type: /qa-smoke storefront

Default TEST_ENV is vcst. Switch with TEST_ENV=vcptcore npm run env:check or TEST_ENV=virtostart ….

Need details? Follow the full setup below.

Prerequisites

Requirement	Version	How to Check	Install
Node.js	18+	`node --version`	nodejs.org
Git	Latest	`git --version`	git-scm.com
IDE	Latest	—	VS Code (recommended), Cursor, or Windsurf

VS Code users: Install the Claude Code extension. Cursor / Windsurf users: MCP support is built-in, no extra extension needed.

Required Accounts & Access

Ask your team lead for:

Anthropic API key — for Claude Code and CI regression
QA environment credentials — frontend + backend URLs, admin and test user accounts (per environment)
GitHub personal access token — for GitHub MCP (code search, PR review)
JIRA (Atlassian) access — VCST project board
Postman API key — for API test collections
Payment processor test cards — Skyflow, CyberSource, Authorize.Net, Datatrance
Figma API key — for design comparison testing (optional)
BrowserStack credentials — for real device / Safari testing (optional)

Step 1: Clone & Install

git clone https://github.com/VirtoCommerce/vc-mcp-testing-module.git
cd vc-mcp-testing-module
npm install
npx playwright install chromium firefox

Edge uses the system-installed msedge — no separate install needed.

Step 2: Environment Variables (Layered Loader)

The project uses a layered env loader keyed by TEST_ENV (default vcst). Files load in order; later files override earlier ones:

File	Tracked	Purpose
`.env.defaults`	git	Cross-env constants (sandbox cards, Builder.io URL)
`.env.${TEST_ENV}`	git	Per-env URLs/identifiers — e.g. `.env.vcst`, `.env.vcptcore`, `.env.virtostart` (no secrets)
`.env.local`	gitignored	Secrets: passwords, API tokens — create this locally
`.env`	gitignored	Legacy fallback (backwards-compat); migrate to `.env.local`

Variable names are stable across environments — only values differ. Agents read via process.env.X; they don't care which file the value came from.

Switching environments

TEST_ENV=vcptcore npm run env:check      # Second QA environment
TEST_ENV=virtostart npm run env:check    # Staging
npm run env:check                        # Default = vcst (current QA)

Required variables

Validate with npm run env:check. 42 variables total; 26 are required (URLs, admin/user credentials, base payment cards, Figma + Postman keys). Optional variables (BrowserStack, 3DS cards, lockout accounts, EUR/multi-org users) fail fast at runtime only for suites that need them.

Create .env.local for secrets:

# ===== Claude / AI =====
ANTHROPIC_API_KEY=sk-ant-your-key-here

# ===== Admin Credentials =====
ADMIN=your_admin_username
ADMIN_PASSWORD=your_admin_password

# ===== Test User Credentials =====
USER_EMAIL=test_user@example.com
USER_PASSWORD=test_password
USER2_EMAIL=second_user@example.com
USER2_PASSWORD=second_user_password
USER_VIRTO=virtostart_user@example.com
USER_VIRTO_PASSWORD=virtostart_password

# ===== Per-env password variants (optional) =====
# Any KEY_${TEST_ENV.toUpperCase()} gets promoted to KEY at runtime.
# Example: lets one .env.local serve all three environments.
USER_PASSWORD_VCPTCORE=different_password_for_vcptcore
USER_PASSWORD_VIRTOSTART=different_password_for_virtostart

# ===== External APIs =====
GIT_TOKEN=ghp_your_github_personal_access_token
FIGMA_API_KEY=your_figma_api_key
POSTMAN_API_KEY=your_postman_api_key
BROWSERSTACK_USERNAME=your_browserstack_username   # optional
BROWSERSTACK_ACCESS_KEY=your_browserstack_access_key  # optional

Per-env URLs, store IDs, and sandbox payment cards live in committed files (.env.defaults, .env.vcst, .env.vcptcore, .env.virtostart) — no setup needed for those.

Accessing env vars in code

import { env } from './config.js';  // ES modules — always use .js extension
console.log(env.FRONT_URL);

Step 3: IDE Setup

VS Code with Claude Code

Install the Claude Code extension from the VS Code marketplace
Open the project folder in VS Code
Set your Anthropic API key when prompted (or via ANTHROPIC_API_KEY in .env.local)
Claude Code automatically reads CLAUDE.md for project context

Cursor / Windsurf

Open the project folder
.mcp.json (created in the next step) configures MCP servers automatically
Agents (.claude/agents/) and skills (.claude/skills/) are tracked in git and load automatically

Step 4: MCP Server Configuration

The .mcp.json file is gitignored — you must create it locally after cloning.

Create `.mcp.json` in the project root

Windows:

{
  "mcpServers": {
    "playwright-chrome": {
      "type": "stdio",
      "command": "cmd",
      "args": ["/c", "npx", "@playwright/mcp@latest", "--config", "config/mcp-playwright-chrome.config.json"]
    },
    "playwright-firefox": {
      "type": "stdio",
      "command": "cmd",
      "args": ["/c", "npx", "@playwright/mcp@latest", "--config", "config/mcp-playwright-firefox.config.json"]
    },
    "playwright-edge": {
      "type": "stdio",
      "command": "cmd",
      "args": ["/c", "npx", "@playwright/mcp@latest", "--config", "config/mcp-playwright-edge.config.json"]
    },
    "postman": {
      "type": "stdio",
      "command": "cmd",
      "args": ["/c", "npx", "@postman/postman-mcp-server@latest", "--minimal"],
      "env": { "POSTMAN_API_KEY": "%POSTMAN_API_KEY%" }
    }
  }
}

macOS / Linux:

{
  "mcpServers": {
    "playwright-chrome": {
      "type": "stdio",
      "command": "npx",
      "args": ["@playwright/mcp@latest", "--config", "config/mcp-playwright-chrome.config.json"]
    },
    "playwright-firefox": {
      "type": "stdio",
      "command": "npx",
      "args": ["@playwright/mcp@latest", "--config", "config/mcp-playwright-firefox.config.json"]
    },
    "playwright-edge": {
      "type": "stdio",
      "command": "npx",
      "args": ["@playwright/mcp@latest", "--config", "config/mcp-playwright-edge.config.json"]
    },
    "postman": {
      "type": "stdio",
      "command": "npx",
      "args": ["@postman/postman-mcp-server@latest", "--minimal"],
      "env": { "POSTMAN_API_KEY": "$POSTMAN_API_KEY" }
    }
  }
}

WebKit is NOT supported on Windows. The repository uses Chromium, Firefox, and Edge.

Browser config files in config/ are tracked in git (viewport 1920x1080, locale en-US, HAR capture enabled, video on failure, isolated contexts).

Additional MCP Servers (user-level, optional)

These are configured in your IDE settings, not in the project .mcp.json:

Server	Purpose
Chrome DevTools MCP	Console logs, network requests, performance tracing, HAR export
Azure MCP	Application Insights, resource health, monitoring
Atlassian MCP	JIRA integration — ticket management, bug reporting (requires Atlassian API token)
Figma MCP	Visual comparison against design specs
GitHub MCP	PR review, code search, issue management (uses `GIT_TOKEN` from `.env.local`)
Context7	Up-to-date library documentation lookup

Step 5: Verify Your Setup

node --version          # Should be 18+
npm run env:check       # Reports SET/EMPTY for all variables — fails if required ones missing

Then open Claude Code in your IDE and type:

Navigate to the storefront URL and take a screenshot

If a browser opens and navigates to the site — you're all set.

Setup Checklist

npm install completed
.env.local created with secrets, npm run env:check passes
.mcp.json created for your OS (Windows: cmd /c npx, macOS/Linux: npx)
Playwright browsers installed (npx playwright install chromium firefox)
IDE installed with Claude Code extension
First test run: browser opens, navigates, takes screenshot

Troubleshooting

"Missing required environment variables"

Run npm run env:check. Verify .env.local exists with secrets; per-env URLs come from the committed .env.${TEST_ENV} file. Check TEST_ENV is set correctly (default vcst).

Playwright MCP server not connecting

Verify .mcp.json exists in the project root (it's gitignored — you must create it)
Check config/ has all 3 browser config JSONs (chrome, firefox, edge)
Run npx playwright install chromium firefox
Restart your IDE after any .mcp.json change

Browser not launching

Close all Chrome windows before using playwright-chrome (user data directory conflict)
On Windows: ensure "command": "cmd" and "args": ["/c", "npx", ...]
On macOS/Linux: use "command": "npx" directly
Do NOT attempt WebKit on Windows — use Edge or Chrome instead

`Browser "chromium" is not installed` from playwright-chrome MCP

Revision mismatch between Playwright versions. Run cli.js install from inside the MCP's bundled playwright-core (no Claude Code restart needed).

"Cannot find module './config.js'"

This project uses ES modules ("type": "module"). Always use .js extensions in imports:

import { env } from './config.js';

JIRA/Atlassian MCP not working

Atlassian MCP is configured at the user/IDE level, not in .mcp.json
Verify your Atlassian API token is set
Check that you have access to the VCST JIRA project

How Testing Works

Four Testing Modes

1. Interactive MCP-Driven Testing (Primary)

The everyday workflow — you tell Claude Code what to test in natural language:

/qa-smoke storefront
/qa-test VCST-1234
Use the qa-frontend-expert to verify the checkout flow

Claude Code opens a real browser via Playwright MCP, navigates the site, performs actions, captures HAR files / screenshots / console logs, and generates reports.

2. CI Regression via Claude Agent SDK

ci/run-regression.ts runs headless regression in Docker. It reads CSV suites from regression/suites/, runs up to 3 in parallel (configurable via MAX_PARALLEL), and produces consolidated reports in reports/regression/ci-YYYY-MM-DD/.

Note: The ci/ directory is gitignored. It's available in the CI environment and for team members with CI access. Contact your team lead if you need local CI setup.

3. Autonomous Interactive Regression (Agent Teams)

autonomous-regression-orchestrator creates a team of child agents using the Agent Teams API. Each child gets an isolated browser, fresh auth, and exponential backoff (30s→60s→120s). The orchestrator manages a 3+1 token bucket (3 browser + 1 reporting), tracks failures with retry + browser fallback chain, and produces a consolidated report with quality gate + optional JIRA tickets.

Invoke: /qa-regression critical --autonomous. Results land in results/{RUN_ID}/.

4. Full Test Cycle CI Pipeline

ci/run-full-cycle.ts chains three phases: sync stale test cases (PR / module / diff driven) → review (7-dimension quality analysis) → regression (affected suites). Each phase has independent skip flags. Triggered on PR merge, daily schedule, or manual dispatch.

Invoke: CHANGE_SOURCE="PR #123" npm run ci:cycle.

Prompt Templates

Located in docs/prompts/:

Template	Purpose
`How to test Builder.io.md`	Builder.io, Virto Pages & vc-frontend testing
`story-testing.md`	Story-level testing prompt

Commands, Skills & Agents

Slash Commands (16)

Type /command-name in Claude Code chat. See .claude/rules/skills-commands.md for full argument reference.

Command	Purpose
`/qa-smoke`	Daily smoke test (~15 min, GO/NO-GO verdict)
`/qa-test`	Test a JIRA ticket, feature, or PR
`/qa-regression`	Run regression suites in parallel (plan-driven `sprint`, or any selection group / IDs)
`/qa-status`	Dashboard: run status, JIRA queue, env health, recent bugs
`/qa-bug`	Reproduce, document, and file a JIRA bug
`/qa-design`	Dual Storybook + Storefront BL-UI audit for components; storefront-only for pages/flows
`/qa-exploratory`	Guided exploratory testing session with heuristics
`/qa-env-check`	Validate env vars, endpoints, MCP servers, test infra
`/qa-coverage-generation`	Orchestrated parallel coverage generation pipeline
`/qa-test-lifecycle`	Unified pipeline: sync stale cases + analyze gaps + generate + review + verify (PR, module, diff, suite, domain)
`/qa-test-plan`	Build a sprint test plan from JIRA + merged PRs in the sprint window
`/qa-sync-tests`	(deprecated — redirects to `/qa-test-lifecycle`)
`/qa-verify-fix`	Verify a bug fix with regression checks + JIRA transition
`/qa-seed-data`	Generate test data via Postman MCP / tear down AGENT-TEST-* entities
`/ba-analyze`	Business analysis with GitHub search + live UI
`/ba-stories`	Generate Agile user stories with BDD criteria

Skills (20)

Type /skill-name in Claude Code chat. Organized in .claude/skills/ across 3 categories:

VC Knowledge (1): /vc-docs — documentation lookup via Context7

Testing (10): /qa-storybook, /qa-accessibility, /qa-design, /qa-plan, /qa-checklist, /qa-api, /qa-coverage-gap, /qa-postman, /qa-seed-data, /qa-review-tests

QA Methodology (9): /qa-process, /qa-investigate, /qa-evidence, /qa-defect, /qa-test-design, /qa-test-cases-generator, /qa-risk, /qa-metrics, /qa-sbtm

Agents (14)

14 specialized agents in .claude/agents/ across two teams (plus a shared shared-instructions.md include). Use them by name in chat:

"Use the qa-frontend-expert to verify the checkout flow"
"Use the qa-backend-expert to test the Platform API"
"Use the ba-story-writer to create stories for VCST-1234"

QA Team (10): qa-lead-orchestrator (sonnet), qa-frontend-expert (opus), qa-backend-expert (opus), qa-testing-expert (opus), test-management-specialist (sonnet), ui-ux-expert (sonnet), regression-orchestrator (sonnet), autonomous-regression-orchestrator (sonnet), autonomous-test-runner (sonnet), test-runner-agent (sonnet), plus shared-instructions

BA Team (4): ba-system-analyzer, ba-api-specialist, ba-story-writer, ba-doc-writer (all sonnet)

QA agents use a four-layer prompt architecture — business logic, domain knowledge, skill set, design decisions. 24 shared knowledge files in .claude/agents/knowledge/ cover API auth, business logic, browser quirks, GraphQL schema, products, sitemap, etc.

Each parallel agent uses its own browser — see .claude/rules/agents.md for browser assignments. Max 3 concurrent browser agents.

Available npm Commands

# Setup & validation
npm install                       # Install dependencies
npm run env:check                 # Validate environment variables (active TEST_ENV layer)

# CI regression (Claude Agent SDK)
npm run ci:smoke                  # Smoke selection (suites 042, 078)
npm run ci:critical               # P0 selection (042, 078, 039, 044, 049)
npm run ci:frontend               # All Frontend/* suites
npm run ci:backend                # All Backend/* suites
npm run ci:full                   # Full regression (all suites, $80 budget)
npm run ci:regression             # Run with SUITE_SELECTION env var
npm run ci:notify                 # Teams notification (requires TEAMS_WEBHOOK_URL)

# Full cycle pipeline (sync → review → regression)
npm run ci:cycle                  # Full 3-phase cycle
npm run ci:cycle:pr               # PR-driven (uses $PR_NUMBER)
npm run ci:cycle:sync-only        # Phase 1 only (sync stale cases)
npm run ci:cycle:no-sync          # Skip Phase 1

# Test data seeding (Postman MCP)
npm run seed                      # All profiles
npm run seed:minimal              # Minimal accounts only
npm run seed:catalog              # Catalogs + products + pricing
npm run seed:full                 # Everything
npm run seed:teardown             # Sweep AGENT-TEST-* entities
npm run seed:dry-run              # Catalog dry-run with verbose output

# GraphQL runner + schema management
npm run schema:refresh            # Refresh introspection snapshot
npm run schema:check              # Drift check (CI gate)
npm run graphql:validate          # Run all GraphQL fixtures
npm run graphql:fixtures:validate # Lint fixtures against current schema
npm run graphql:fixtures:update   # Sync fixtures to schema
npm run graphql:lint-labels       # Lint runner case labels
npm run graphql:cleanup-leaked    # Sweep leaked test orgs

# Suite manifest + reports
npm run suites:sync               # Regenerate `selections` from `_doc` rules
npm run suites:lint               # Check selections are in sync
npm run scope:validate            # Validate critical-ui-scope coverage
npm run report:regression         # Generate HTML regression report
npm run report:graphql            # Generate HTML GraphQL report
npm run refresh-product-guids     # Refresh product GUID fixtures

Repository Structure

vc-mcp-testing-module/
├── CLAUDE.md                # Claude Code project instructions
├── .claude/
│   ├── agents/              # 14 agent definitions (qa/ + ba/) + knowledge/ (24 reference files)
│   ├── skills/              # 20 skills across 3 categories
│   ├── commands/            # 16 slash commands
│   └── rules/               # Reference docs (agents, regression, skills-commands, mcp-browsers, test-data)
├── config/                  # Playwright browser configs + test-suites.json manifest
├── ci/                      # CI regression + full-cycle pipeline (gitignored)
├── docs/prompts/            # LLM prompt templates
├── regression/suites/       # 99 CSV suites (~3,756 test cases)
│   ├── Frontend/            # 46 CSVs in 15 module dirs (auth, catalog, cart, …)
│   ├── Backend/             # 52 CSVs in 27 module dirs (platform, api, graphql, …)
│   └── _release/            # Master release suite (080)
├── tests/                   # Test cases by sprint/JIRA ticket
├── test-data/               # Aliases registry + CSV fixtures (orgs, addresses, accounts)
├── reports/                 # Bug reports + regression reports
├── scripts/                 # Resolvers, GraphQL runner, sync/lint utilities, seeders
├── config.js                # Layered env loader (TEST_ENV-keyed)
└── get_variables_env.js     # Environment validation script

Tracked in git: .claude/, config/, regression/, tests/, docs/, scripts/, test-data/, CLAUDE.md, .env.defaults, .env.${TEST_ENV}

Gitignored: .env, .env.local, .env.backup, .mcp.json, settings.json, results/, ci/, .github/, .claude/settings.local.json

Regression Test Suites

99 suites with ~3,756 test cases in enriched agent-native CSV format. Authoritative definitions in config/test-suites.json (_meta.totalSuites: 99).

Suites are organized under Frontend/<module>/ and Backend/<module>/ directories, with IDs like 001-catalog-navigation.csv, 050a-graphql-xcatalog.csv, 080-full-regression-release.csv.

Selection Groups

Selection	Description	Use Case
`smoke`	042, 078	Daily validation before deployment
`critical`	042, 078, 039, 044, 049	P0 suites only
`release`	080	Master release suite (100 cases) for major releases
`purchase-flow`	Cart + checkout + orders + payment	Purchase flow regression
`catalog` / `search` / `orders` / `auth` / `b2c` / `marketing` / `platform`	Module-aligned	Focused module runs
`bopis` / `payment` / `configurable-products` / `whitelabeling`	Feature-aligned	Focused feature runs
`layout-stability`	048b	Critical-UI-scope regression (7 components × 8 pages matrix)
`frontend`	All Frontend/ suites	Frontend-only regression
`backend`	All Backend/ suites	Backend-only regression
`sprint`	Plan-driven via `vc/shared/docs/Sprint plans/sprint-*-summary.json` (`--no-plan` falls back to P0+P1)	Before sprint release
`sprint:XX-YY`	Pinned to a specific sprint plan	Re-run a past sprint's scope
`full`	All 99 suites	Before production release

P0 Critical Suites

Suite	Domain
042 — Smoke Tests	Core flows end-to-end
078 — Smoke companion	Companion to suite 042
039 — Payment (CyberSource)	Embedded payment form, 3DS
044 — Security Tests	PCI compliance, input validation, XSS, CSRF
049 — Platform REST API	REST API health

CSV Column Format

ID, Title, Section, Priority, Business_Rule, Edge_Case_Refs, Preconditions,
Test_Data, Steps, Assertions, Cross_Layer_Checks, Failure_Signals, Cleanup,
References, Automation_Status

Authoring guides:

Browser-mode CSV tags: .claude/agents/knowledge/test-runner-tags.md
Runner-native GraphQL: .claude/agents/knowledge/graphql-test-cases-runner.md
Test data resolver (@td() + {{VAR}}): .claude/rules/test-data.md

Frontend / Backend module layout

Frontend (46 CSVs): auth/, catalog/, search/, cart/, checkout/, orders/, payment/, bopis/, b2c/, configurable-products/, whitelabeling/, marketing/, loyalty/, cross-cutting/, smoke/

Backend (52 CSVs): platform/, store/, catalog/, customer/, pricing/, inventory/, marketing/, notifications/, cms/, orders/, api/, graphql/, search/, configurable-products/, whitelabeling/, assets/, channels/, contracts/, image-tools/, import-export/, loyalty/, push-messages/, returns/, seo/, shipping/, xmarketing/, smoke/

Full suite-to-module mapping lives in .claude/agents/knowledge/module-suite-map.md.

CI/CD Pipeline

Note: The ci/ directory is gitignored. CI infrastructure is available in the GitHub Actions environment. Contact your team lead for local Docker CI setup.

GitHub Actions

Go to Actions > Regression Tests > Run workflow:

Parameter	Options	Default
Suite selection	`smoke`, `critical`, `sprint`, `full`, `frontend`, `backend`, or comma-separated IDs	`smoke`
Environment	`qa`, `staging`	`qa`
Max budget (USD)	Any number	`10.0`

CI environment mapping: qa → FRONT_URL / BACK_URL, staging → VIRTO_START_FRONT / VIRTO_START_BACK. CI mode uses only playwright-chrome (single headless Chromium) for all suites.

Scheduled Runs

Daily smoke: Mon-Fri 6:00 AM UTC — suite 042, $5 budget
Weekly full: Sunday 2:00 AM UTC — all 99 suites, $80 budget
Full-cycle pipeline (.github/workflows/full-cycle.yml): triggered on PR merge to main; daily Mon-Fri 8 AM UTC; or manual dispatch

Cost Estimates

Selection	Suites	Est. Cost	Est. Time
`smoke`	2	~$2-5	~30 min
`critical`	5	~$10-15	~2 hrs
`sprint`	varies (plan-driven)	~$40-60	~5-8 hrs
`full`	99	~$60-100	~10-14 hrs

Required GitHub Secrets

Secret	Required	Description
`ANTHROPIC_API_KEY`	Yes	Claude API key
`FRONT_URL` / `BACK_URL`	Yes	QA environment URLs
`ADMIN` / `ADMIN_PASSWORD`	Yes	Admin credentials
`USER_EMAIL` / `USER_PASSWORD`	Yes	Test user credentials
`USER2_EMAIL` / `USER2_PASSWORD`	Yes	Second test user
`USER_VIRTO` / `USER_VIRTO_PASSWORD`	Yes	Virtostart user
`STORE_ID`	Yes	Store identifier
`STORYBOOK_URL` / `STORYBOOK_DEV_URL`	Yes	Storybook URLs
`VIRTO_START_FRONT` / `VIRTO_START_BACK`	For staging runs	Virtostart URLs
`SKYFLOW_`, `CYBERSOURCE_`, `AUTHORIZNET_`, `DATATRANCE_`	For payment suites	Payment test cards
`POSTMAN_API_KEY` / `FIGMA_API_KEY`	Yes	API integrations
`GIT_TOKEN`	Yes	GitHub PAT for GitHub MCP
`TEAMS_WEBHOOK_URL`	Optional	Teams notification webhook

Testing Domains & Browser Matrix

Critical Revenue Flows (must pass before deployment)

Registration/Auth, Catalog/Facets, Cart (variations, BOPIS), Search, Addresses, Checkout/Payment, Orders, B2B Multi-org, GA4 tracking.

Payment flow: CyberSource shows the payment form on the cart page. All other processors (Skyflow, Authorize.Net, Datatrance) require clicking "Place Order" first → redirects to /checkout/payment.

Browser Matrix

Browser	Engine	MCP Server	Priority
Chrome	Chromium	`playwright-chrome`	Primary
Edge	Chromium (msedge channel)	`playwright-edge`	High
Firefox	Gecko	`playwright-firefox`	High

WebKit/Safari is NOT supported on Windows. For Safari testing, use BrowserStack.

Mobile testing via BrowserStack (BROWSERSTACK_USERNAME / BROWSERSTACK_ACCESS_KEY).

Theme: Coffee (only Coffee passes WCAG A11y tests). Communication: Microsoft Teams.

Resources

Playwright MCP
Claude Agent SDK
Virto Commerce Documentation
Storefront User Guide
Project Issues
Detailed internal reference: .claude/rules/ (agents, regression, skills-commands, mcp-browsers, test-data)

Name		Name	Last commit message	Last commit date
Latest commit History 492 Commits
.claude-plugin		.claude-plugin
.claude		.claude
.github/workflows		.github/workflows
.newman-run		.newman-run
bootstrap		bootstrap
ci		ci
config		config
docs		docs
regression/suites		regression/suites
reports		reports
scripts		scripts
templates		templates
test-data		test-data
tests/Sprint26-11		tests/Sprint26-11
vc		vc
.env.defaults		.env.defaults
.env.localhost		.env.localhost
.env.vcptcore		.env.vcptcore
.env.vcst		.env.vcst
.env.virtostart		.env.virtostart
.gitattributes		.gitattributes
.gitignore		.gitignore
CHANGELOG.md		CHANGELOG.md
CLAUDE.md		CLAUDE.md
INDEX.md		INDEX.md
README.md		README.md
config.js		config.js
get_variables_env.js		get_variables_env.js
manifest.json		manifest.json
package-lock.json		package-lock.json
package.json		package.json

Folders and files

Latest commit

History

Repository files navigation

vc-mcp-testing-module

Quick Start

Table of Contents

Prerequisites

Required Accounts & Access

Step 1: Clone & Install

Step 2: Environment Variables (Layered Loader)

Switching environments

Required variables

Accessing env vars in code

Step 3: IDE Setup

VS Code with Claude Code

Cursor / Windsurf

Step 4: MCP Server Configuration

Create .mcp.json in the project root

Additional MCP Servers (user-level, optional)

Step 5: Verify Your Setup

Setup Checklist

Troubleshooting

"Missing required environment variables"

Playwright MCP server not connecting

Browser not launching

Browser "chromium" is not installed from playwright-chrome MCP

"Cannot find module './config.js'"

JIRA/Atlassian MCP not working

How Testing Works

Four Testing Modes

1. Interactive MCP-Driven Testing (Primary)

2. CI Regression via Claude Agent SDK

3. Autonomous Interactive Regression (Agent Teams)

4. Full Test Cycle CI Pipeline

Prompt Templates

Commands, Skills & Agents

Slash Commands (16)

Skills (20)

Agents (14)

Available npm Commands

Repository Structure

Regression Test Suites

Selection Groups

P0 Critical Suites

CSV Column Format

Frontend / Backend module layout

CI/CD Pipeline

GitHub Actions

Scheduled Runs

Cost Estimates

Required GitHub Secrets

Testing Domains & Browser Matrix

Critical Revenue Flows (must pass before deployment)

Browser Matrix

Resources

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Create `.mcp.json` in the project root

`Browser "chromium" is not installed` from playwright-chrome MCP

Packages