[BUG] Server is temporarily limiting requests breaks workflows

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?

Whenever anthropic decides to limit the number of requests per user (API Error: Server is temporarily limiting requests), it completely breaks running ultracode workflows. They skip the broken subagents instead of retrying and continuing them, losing all hundreds of thousands worth context consumed in those subagents and burning millions of tokens in a now useless end result because half the subagents were silently dropped.

What Should Happen?

Claude workflows should retry failed subagents, especially when claude arbitrarily chose to make them fail in the first place as a form of managing consumption. Managing consumption by making 75% of users' workflows fail part way thru, and then making the user front the bill for that, is not acceptable.

Error Messages/Logs

API Error: Server is temporarily limiting requests

Steps to Reproduce

Create an ultracode workflow while anthropic has rate limiting turned on, or create a workflow and then turn on the rate limiting part way thru. Watch most subagents fail, and the workflow continue as if they hadn't, and finally watch all your spent money and tokens burn.

Claude Model

None

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

V2.1.183

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

Windows Terminal

Additional Information

_No response_