[BUG] Core CLI tools (Bash/Read/Write/Edit) silently missing during backend degradation — May 18-19 2026

Resolved 💬 2 comments Opened May 19, 2026 by hc-7000 Closed Jun 18, 2026

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?

What happened

During Anthropic's backend degradation on May 18-19 2026, all core CLI tools (Bash, Read, Write, Edit, Grep, Glob) silently disappeared from Claude Code sessions with zero warning or error message.

  • Sessions started normally
  • Claude appeared to work (responded to messages)
  • But core tools were completely unavailable — ToolSearch returned no results for Bash, Read, Write, Edit, Grep, Glob
  • No error, no degraded-mode warning, nothing in the CLI to indicate the backend was broken
  • Users spent hours troubleshooting, thinking it was a local config issue

Anthropic's own status page (status.claude.com) recorded ~525 minutes of partial outages on May 18 across Claude Code, API, and claude.ai (elevated errors on Opus 4.6 and Haiku 4.5). Issues continued May 19 with elevated login errors and tool failures.

Related bug filed same day: #60449 (Bash tool runs in background mode, output never returned, macOS)

Steps to reproduce

  1. Start a new Claude Code session during a backend degradation event
  2. Attempt to use Bash, Read, Write, Edit, Grep, or Glob tools
  3. Tools are silently unavailable — no error, no warning

Environment

  • Platform: macOS
  • Claude Code version: latest (v2.1.x)
  • Model: claude-opus-4-6
  • Date: May 18-19, 2026

Impact

Users had no way to know the backend was degraded. They wasted significant time fighting what appeared to be broken local configuration, trying model switches, workarounds, and session restarts — all because there is zero visibility into backend health from inside the CLI.

This is directly related to closed issue #33328 (Real-time backend health/trust indicator in CLI) which was closed as "Not Planned". This incident is exactly the scenario that feature would have prevented.

What Should Happen?

Claude Code should display a clear warning when core tools are unavailable due to backend degradation, rather than silently failing. A status indicator showing backend health would prevent users from wasting time on false troubleshooting.

Error Messages/Logs

Steps to Reproduce

Here's what I put in the steps to reproduce section:

  1. Start a new Claude Code session during a backend degradation event
  2. Attempt to use Bash, Read, Write, Edit, Grep, or Glob tools
  3. Tools are silently unavailable — no error, no warning

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

2.1.143

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗