Pod

Yes, this is on purpose. Pod is built for agents, so the default page is plain HTML.
Human? View the normal website.
Agent? You probably prefer /mcp/mcp-2/issues.md or /mcp/mcp-2/issues.json, or Pod over MCP.

Reported issues for mcp

Pod holds 7 of 7 problems reported by people outside the maintainer team. Issues filed by the project's own owners, members and collaborators are excluded entirely — a maintainer's release checklist is not a warning to a prospective user.

Back to mcp.

Most discussed

[Bug]: Sauce Labs provider discards user-supplied options and injects an empty tunnel name

Environment

Summary

The Sauce Labs provider's buildCapabilities() builds its own sauce:options object and assigns it to the resolved capabilities. Every other top-level user capability is merged via spread, but sauce:options is a nested object the provider also constructs, so it gets overwritten wholesale — any Sauce settings a calle

Read the thread · 2026-07-20 · closed · outside contributor · 2 comments

Attach to an existing Appium session via sessionId

Summary

Currently, there's no way to attach to an existing Appium session.

start_session can attach to a locally-launched Chrome over CDP (added in v2.4.0, which closed #31), but there's still no way to attach to an existing remote WebDriver / Appium session by its sessionId – e.g. an already-running BrowserStack or Sauce Labs cloud session, or an Appium native-app session that was provisioned outside the MCP.

Why the existing attach doesn't cover it

The v2.4.0 attach pat

Read the thread · 2026-07-14 · closed · outside contributor · 1 comment

[Bug] v3.5.0: require("../../package.json") resolves to wrong path in scoped package — crashes on startup

Bug Report

Environment

Description

Starting with v3.5.0, the MCP server crashes immediately on startup with Process exited with code 1 when launched via node lib/server.js from the global (or local node_modules) installation.

Root Cause

In the bundled lib/server.js (around line 2133-2135), the code uses:

var require2 = createRequire(import.meta.url);
var package_default = require2("..

[Read the thread](https://github.com/webdriverio/mcp/issues/112) · 2026-05-21 · closed · external user · 1 comment

### Feature Request: Native LambdaTest Support as a Cloud Provider

## Summary
 
Add native LambdaTest support to `wdio-mcp` as a cloud provider, following the same pattern as the existing BrowserStack integration. This would allow users to start mobile and browser sessions on LambdaTest real devices without exposing credentials through the AI agent context.
 
---
 
## Motivation
 
Currently, `wdio-mcp` supports BrowserStack as a cloud provider via:
 
```json
{
  "provider": "browserstack",
  "platform": "android"
}

The MCP server reads `BROWSERSTACK_USERN

Read the thread · 2026-05-11 · closed · external user · 4 comments

[bug] ]start_session hardcodes protocol: "http" — breaks HTTPS cloud providers (e.g. LambdaTest)

Bug Report

Description

start_session hardcodes protocol: "http" when building the Appium connection config in server.js. This breaks any cloud provider that uses HTTPS (port 443), such as LambdaTest's real device hub.

Steps to Reproduce

  1. Configure .mcp.json with LambdaTest credentials
  2. Call start_session with raw capabilities pointing to LambdaTest:
    • hostname: "mobile-hub.lambdatest.com"
    • port: 443
    • protocol: "https"
  3. Session fails — the protoco

Read the thread · 2026-04-24 · closed · external user · 2 comments

Attach an existing session to MCP

There should be a way to attach the existing session to MCP. It seems start_new_session command alway start a new session. attaching an existing session can help to run the test from script/command line i.e. WDIO test runner and then handover to AI assistance or agent.

Read the thread · 2026-02-18 · closed · external user · 5 comments

feat(start_browser): allow passing browser capabilities (extensions, user-data-dir, chrome args) + optional attach to existing Chrome

Summary

I’m using the WebdriverIO MCP start_browser tool to automate end-to-end flows that require launching the browser with custom capabilities, especially Chrome extensions (any extension) and/or a persistent profile.

Right now, start_browser only exposes basic launch parameters (headless/window size/navigationUrl). There’s no way to pass WebDriver capabilities such as goog:chromeOptions (e.g. --load-extension, --user-data-dir, custom args / prefs).

**Concrete examp

Read the thread · 2026-02-17 · closed · outside contributor · 2 comments

Most recent

The remaining reports are on the project's issue tracker.