[BUG] FleetView dispatched agents doesn't properly use the correct version when dispatching an agent with an up to date version while an existing previous version is running on the background

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 dispatching an agents with the new agents view with an up to date claude code in a terminal while a process using a previous version is still active, the claude code session with latest version is unable to dispatch an agent without an ENOENT error

What Should Happen?

Claude agents view should dispatch agents without errors

Error Messages/Logs

[worker crashed (ENOENT: no such file or directory, posix_spawn '/opt/homebrew/Caskroom/claude-code@latest/2.1.139/claude') — respawning…]

Steps to Reproduce

1. Install Claude Code via Homebrew cask, launch a session.
2. While the session is running, let the cask auto-upgrade (or run brew upgrade --cask claude-code@latest).
3. Dispatch any agent / FleetView worker — posix_spawn ENOENTs on the old versioned path even though /opt/homebrew/bin/claude already points to the new version.
4. Quitting the UI doesn't fix it if any host process survives (e.g. --continue reattaches to the same poisoned host).

Likely root cause on their side: the parent caches process.execPath at startup and reuses it to spawn workers. Fix is to either re-resolve /opt/homebrew/bin/claude (or the PATH-found claude) at spawn time, or detect ENOENT on spawn and retry through PATH.

Similar bug #18019 and #25364

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

2.1.139

Claude Code Version

2.1.140

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Other

Additional Information

_No response_