[BUG] Subagents can no longer prompt me for permission

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?

It seems like Claude Code now launches any/all subagents as background tasks, no matter what. Likewise, subagents are no longer able to ask me for permission to do things, like:

• If I just cloned a git repo into a temporary directory (like into $(mktemp -d)) and ask it to go explore it
• If a web fetch is performed on an untrusted URL

There doesn't seem to be a way around this without just asking the parent process to read/fetch the relevant data first.

What Should Happen?

Like with synchronous subagents, asynchronous agents show allow prompting for permissions. At the least, this should be a toggle when the agent spins up.

Error Messages/Logs

Steps to Reproduce

Ask Claude Code to spin up a subagent to fetch a URL it hasn't seen before.

Claude Model

Not sure / Multiple models

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

2.1.179

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

Other

Additional Information

_No response_