claude-code - 💡(How to fix) Fix /mcp TUI locks terminal — no keys respond, must kill shell

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…

Root Cause

MCP setup

  • ~16 MCP servers configured: 6 healthy local/SSE, ~10 organization-managed claude.ai * connectors that show "! Needs authentication" or "✗ Failed to connect" in claude mcp list.
  • The org-managed connectors cannot be removed from the CLI (claude mcp remove reports "no server found") and appear to be greyed-out / non-removable in claude.ai → Settings → Connectors, presumably because they're pushed by an org admin.

Fix Action

Fix / Workaround

Attempted workarounds

  • export MCP_TIMEOUT=3000 — no effect (it controls server startup timeout, not the TUI).
  • Removing the one local OAuth-pending server (sentry) — no effect; the org-managed claude.ai Sentry remains and still hangs /mcp.
RAW_BUFFERClick to expand / collapse

Bug

Running /mcp in the Claude Code CLI hangs the terminal immediately. The interactive picker normally lets me arrow up/down between MCP servers, but in this state no input works — arrow keys, Esc, Ctrl+C, nothing. Only recourse is killing the shell.

Environment

  • Claude Code: 2.1.118
  • macOS: 26.5 (build 25F5042g)
  • Shell: zsh

MCP setup

  • ~16 MCP servers configured: 6 healthy local/SSE, ~10 organization-managed claude.ai * connectors that show "! Needs authentication" or "✗ Failed to connect" in claude mcp list.
  • The org-managed connectors cannot be removed from the CLI (claude mcp remove reports "no server found") and appear to be greyed-out / non-removable in claude.ai → Settings → Connectors, presumably because they're pushed by an org admin.

Steps to reproduce

  1. Have several claude.ai * connectors in unauthenticated / failed state (visible in claude mcp list).
  2. Launch claude.
  3. Type /mcp.

Expected

  • TUI renders, arrow keys navigate the server list, Esc exits.

Actual

  • TUI does not render usable state. All keyboard input is swallowed. Must kill the terminal process to recover.

Attempted workarounds

  • export MCP_TIMEOUT=3000 — no effect (it controls server startup timeout, not the TUI).
  • Removing the one local OAuth-pending server (sentry) — no effect; the org-managed claude.ai Sentry remains and still hangs /mcp.

Suggested fixes

  • /mcp should be cancellable (Esc / Ctrl+C should always work).
  • /mcp should render with a timeout per server rather than blocking the entire TUI until all servers respond.
  • Provide a way to opt out of org-pushed connectors locally (~/.claude/settings.json flag or per-server disable).

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING