[BUG] Claude cannot see failed/erroring MCP tools — no visibility into configured-but-broken MCPs

Open 💬 1 comment Opened Jun 30, 2026 by marek-orgon

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

When an MCP tool is configured but fails to connect/load, Claude has zero visibility into it. From Claude's perspective, the tool simply doesn't exist — it cannot distinguish between "not configured" and "configured but broken".

Claude tells the user the tool is not available, when in fact it is configured but failing. The user has to manually inform Claude of something the system already knows.

What Should Happen?

Claude should receive system-level context listing all configured MCP tools and their connection status, e.g.:

MCP tools status:

  • some-tool: failed to connect
  • working-tool: connected

So Claude can proactively say "X is configured but failing — check your credentials/endpoint" instead of incorrectly claiming it has no access.

Error Messages/Logs

Steps to Reproduce

  1. Configure an MCP tool in Claude Code settings (e.g. via mcp.json)
  2. Make the tool fail to connect (wrong credentials, unreachable endpoint, etc.)
  3. Ask Claude to use that tool
  4. Claude responds that the tool is not available / not installed
  5. Claude has no idea the tool is configured but broken

Claude Model

Sonnet (default)

Is this a regression?

No, this never worked

Last Working Version

_No response_

Claude Code Version

2.1.185 (Claude Code)

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗