QE MCP Server Guide

What our MCP servers do, when to use them, and how to add them in Cursor or VS Code.

What is MCP?

Model Context Protocol (MCP) is a standard way for editors like Cursor and VS Code to connect to tools over the network. Our QE MCP servers are internal Quality Engineering endpoints that help developers generate, review, and improve tests using shared INDG quality standards.

These servers work alongside Cursor rules and IDE agent workflows so generated tests follow the patterns our teams expect for Playwright, QE Blocks, unit testing, accessibility checks, test gap analysis, test planning (Test Spec), and browser automation.

Most V2 tools return meta-prompts that an agent executes in your target repository. Many write working artifacts under .qe/reports/ (add .qe/ to .gitignore). Typical flow: plan with Test Spec Generator → implement with E2E / Unit / Blocks V2 → analyze with Test Gap Analyzer → heal failures with playwright_test_analyzer → audit a11y with Accessibility Audit.

MCP and Cursor rules help standardize how we work and reduce rework, but they do not replace developer judgment. Developers are still responsible for reviewing generated output, validating user intent, and owning what ships to production.

Available MCP servers

MCP mounts expose QE capabilities to Cursor and VS Code. The table below is populated from the production catalog at https://qe-mcp.bindg.com/tools. Each row shows the mount URL to add in your IDE; where a dedicated guide exists, use the Guide link for setup prompts and examples.

Loading MCP catalog…

Name	What it does	Mount URL	Guide
Loading…

Catalog API: GET https://qe-mcp.bindg.com/tools (dev: https://qe-mcp-dev.bindg.com/tools). OpenAPI: qe-mcp.bindg.com/docs.

Why these servers exist

Quality Engineering teams often lose time and consistency due to:

Manual processes for test planning, code generation, coverage, and gap analysis.
Inconsistent test structure, documentation, and assessment frameworks across teams.
Limited time for maturity assessments and strategic planning.
Difficulty identifying missing test cases between app code and automation.
Knowledge silos around best practices and reusable components.

How they improve our process

Automation

Reduces repetitive work from hours to minutes.

Standardization

Encourages consistent test structure, naming, patterns, and assessment frameworks.

Efficiency

Speeds up test generation, test planning, and repo analysis.

Intelligence

Supports systematic gap analysis to identify missing cases and likely defect areas.

Accessibility of knowledge

Centralizes best practices, reusable patterns, and recommended components.

What you can do with them

Author requirements and test criteria, then generate plan + agent prompt packs (Test Spec).
Generate Playwright E2E tests for core user journeys using org conventions (E2E V2).
Heal failing Playwright tests with playwright_test_analyzer.
Use the Playwright browser MCP for live UI inspection (not primary test generation).
Generate Playwright tests that use QE Blocks, or transform plain Playwright to Blocks.
Generate unit tests, coverage recommendations, and mutation-testing feedback.
Identify test gaps, mitigate risk, score test quality, and find suite redundancy.
Apply consistent patterns such as AAA, TDD, POM, and Screenplay.
Support accessibility audits and fixes through dedicated accessibility tools.
Support API and k6 performance testing through E2E V2 sibling tools.

Adding MCP servers to your IDE

The only structural difference between the two editors is that Cursor uses mcpServers with transport, and VS Code uses servers with type. URLs and server names are the same.

Cursor — `~/.cursor/mcp.json`

{
  "mcpServers": {
    "qe_e2e_tests_v2": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/e2e_tests",
      "transport": "http"
    },
    "qe_blocks_tests_v2": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/blocks",
      "transport": "http"
    },
    "qe_unit_tests_v2": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/unit_tests",
      "transport": "http"
    },
    "test_gap_analyzer": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/test_gap_analyzer",
      "transport": "http"
    },
    "test_spec_generator": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/test_spec_generator",
      "transport": "http"
    },
    "accessibility_audit": {
      "url": "https://qe-mcp.bindg.com/mcp/accessibility_audit",
      "transport": "http"
    },
    "playwright": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/playwright",
      "transport": "http"
    }
  }
}

VS Code — `.vscode/mcp.json`

{
  "servers": {
    "qe_e2e_tests_v2": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/e2e_tests",
      "type": "http"
    },
    "qe_blocks_tests_v2": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/blocks",
      "type": "http"
    },
    "qe_unit_tests_v2": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/unit_tests",
      "type": "http"
    },
    "test_gap_analyzer": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/test_gap_analyzer",
      "type": "http"
    },
    "test_spec_generator": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/test_spec_generator",
      "type": "http"
    },
    "accessibility_audit": {
      "url": "https://qe-mcp.bindg.com/mcp/accessibility_audit",
      "type": "http"
    },
    "playwright": {
      "url": "https://qe-mcp.bindg.com/mcp/v2/playwright",
      "type": "http"
    }
  },
  "inputs": []
}

Add to Cursor

Open Cursor and go to Settings.
Search for Tools & MCP.
Choose New MCP Server.
Paste the JSON above (or just one entry).
Save.
Confirm the server appears in Installed MCP Servers.
Start the server if it is not running, then run a smoke-test prompt.

Add to VS Code

Open the Command Palette (Cmd+Shift+P on Mac, Ctrl+Shift+P on Windows).
Search for MCP: Add Server.
Select the HTTP option.
Enter the MCP server URL (for example https://qe-mcp.bindg.com/mcp/v2/playwright).
Provide an alias (for example playwright).
Choose whether to save globally or to the workspace.
Open the installed MCP server list and start the server.
Confirm the server is connected.

For internal setup notes, see Adding QE's MCP Servers to your IDE.

BBVPN proxy configuration

If you are connected via BBVPN, configure your IDE to bypass the corporate proxy for qe-mcp.bindg.com. Without this, VS Code or Cursor may fail to reach https://qe-mcp.bindg.com.

The http.noProxy setting is not MCP server configuration — it lives in your editor settings, not in mcp.json.

Cursor / VS Code settings JSON

{
  "http.noProxy": [
    "qe-mcp.bindg.com"
  ]
}

On macOS, the Cursor/VS Code settings file is typically at ~/Library/Application Support/Code/User/settings.json for VS Code and ~/.cursor/settings.json (or the Settings UI) for Cursor. You can also use the Settings UI: search for proxy, then add qe-mcp.bindg.com to the No Proxy list.

Validate the connection

After adding a server, verify it is connected and started. Use a simple smoke-test prompt in an agent chat, for example:

Smoke test: Using my MCP server playwright, list available tools or capabilities and provide a short ready checklist.

E2E architecture check: Using my MCP server qe_e2e_tests_v2, analyze this repo and propose a Playwright E2E test architecture.

Unit test plan: Using my MCP server qe_unit_tests_v2, generate a unit test plan for the core business logic modules.

Test Spec: Using my MCP server test_spec_generator, help me write a test plan and criteria for the login feature.

If the MCP server does not respond, confirm that:

You are connected to the required network/VPN.
BBVPN proxy bypass is configured if required.
The server URL is correct.
The server is started in your IDE.
The JSON uses the correct structure for your IDE (mcpServers/transport vs servers/type).

External references

MCP server catalog: qe-mcp.bindg.com/docs
Adding MCP servers to your IDE (Confluence): Adding QE's MCP Servers to your IDE
QE MCP Servers overview (Confluence): QE MCP Servers — Overview

← Back to Quality Engineering Tools Suite