[BUG] Claude Code spawns processes in tight loop, saturates CPU and causes system-wide fork failed: resource temporarily unavailable

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?

Environment:

• Claude Code: 2.1.37
• Node: 20.19.6 (via Volta)
• macOS: 26.2 (25C56), Apple Silicon (ARM64)
• RAM: 48GB

Problem:
While Claude Code waits on a long-running child process (Hive query
via shell script), the node process spikes to 95-98% CPU and makes
the entire system unforkable — even ls fails with
zsh: fork failed: resource temporarily unavailable.

Process/thread counts don't change significantly, and memory isn't
the issue. The system has ~14GB free.

Root cause from sample output:
The stack trace shows Array.forEach calling ProcessWrap::Spawn →
posix_spawn in a tight loop with no concurrency control. 3,730 of
3,903 samples (95.6%) are stuck in posix_spawn. Claude Code appears
to be rapidly spawning processes without throttling or backpressure.

Impact:
System becomes completely unusable — cannot fork any new process
from any terminal. Only fix is force-killing the Claude node process,
which itself requires Activity Monitor since shell commands can't fork.

claude_sample.txt

What Should Happen?

It shouldn't spawn processes like that

Error Messages/Logs

Steps to Reproduce

I ran claude code, then asked it to execute a shell script to run a hive query. But I have seen the same thing happening right after starting claude code and typing "/" character to list commands. It just happens at random times, and something like every 2nd session

Claude Model

Opus

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.37

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

iTerm2

Additional Information

_No response_