›_ Claude Code Issues
Claude Code Issues / Error strings / MCP server failed to start or connect / Page 3

MCP server failed to start or connect — page 3

A configured MCP server never completed its handshake or dropped out.

230 issues · 46 open · 184 resolved (80%) · first seen Mar 5, 2025

First response · integration

What to do now

Most likely: A local integration or child process did not satisfy Claude Code’s expected contract.

  1. Isolate the failing MCP server, shell command or editor integration.
  2. Run that dependency directly and capture its exit code and stderr.
  3. Compare with a clean configuration before editing global settings.

These are conservative triage steps, not an official Anthropic fix. Use the issue and workaround evidence below before making a destructive configuration change.

Is this getting better or worse?

This class of problem is converging. 61 new reports in the last 90 days vs 89 in the 90 before — -31%. The open backlog peaked at 46 in 2026-08 and sits at 46 today.

0 12 23 35 46 2025-032025-092026-032026-08
2025-03 — First reports2026-04 — Most new reports (34)2026-08 — Peak backlog (46 open)

Derived by counting each matching issue from its creation date until it closed. First-party computation over the anthropics/claude-code tracker — the trend is not published anywhere upstream.

Which builds report this

137 of these reports name the Claude Code build they were running, spanning 85 releases. Heaviest reporters:

Report volume, not a defect count: an issue can mention a build without that build having caused the failure. Versions come from the version string in the bug-report template, matched against real published tags.

Do these get fixed?

Median time to close is 27 days across 184 closures. Of the 184 closures with a recorded reason, 21% were closed as completed and 146 as not-planned or duplicate. The most recent completed fix landed in 2026-08.

Workarounds reported by the community

78 of these issues post a workaround someone says worked. The highest-engagement ones:

Claude.ai MCP servers auto-injected into Claude Code without opt-in — causes OOM crashes on resource-constrained systems
…r's Claude.ai account and load them into every CLI session. Workaround Manually editing to set: However, this is a cached flag and may be overwritten on next update/startup. Expected Behavior 1. This should be opt-in, no…
Found in the description of #20412 · resolved
[BUG] Slack plugin fails to authenticate - "does not support dynamic client registration"
…blocking real workflow automation for us. Why the bot token workaround ( ) doesn't work: - Bot tokens require the bot to be invited to every channel. In an enterprise workspace with dozens of relevant channels, this is i…
Found in the thread of #18009 · resolved
[FEATURE] Claude can reconnect to a broken MCP in a manner similar to how it performs other actions
…call pauses the conversation rather than letting the model work around it. Together with the reconnection UX proposed in this issue, the behavior tiers would be: Scenario Action ---------- -------- Server process crashe…
Found in the thread of #10071 · still open
[BUG] Claude Code SDK: SDK MCP server faills to connect due to closed stream
I encountered the same issue today and was able to work around it with a minimal adjustment: 1. Refactored into an MCP Server using (this was straightforward since the code format is nearly identical). 2. Updated the def…
Found in the thread of #6710 · resolved
[BUG] OAuth Authentication Succeeds but MCP Reconnection Fails - Requires Restart
…ansport - Server is properly configured for streamable-http Workaround Restart Claude Code after OAuth authentication completes. The stored tokens are preserved and work correctly after restart. Impact - Poor user experi…
Found in the description of #10250 · resolved

Excerpts are quoted from the linked issue so you can judge relevance before opening it. Threads that only ask for a workaround, or report that there is none, are excluded. These are community suggestions, not official guidance from Anthropic — read the full thread before applying one.

Matching issues 201–230

Ranked by community engagement (reactions weighted over comments).

Related aggregations

  1. Garbled / flickering terminal output

    831 issues

  2. spawn ENOENT / command not found

    506 issues

  3. API Error 400 (invalid_request_error)

    2,591 issues

  4. AbortError: the operation was aborted

    2,101 issues

  5. ENOENT: no such file or directory

    1,200 issues

  6. API Error 500 (internal server error)

    603 issues