[BUG] 16% of 5-hour usage limit consumed by simple investigation prompt.

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?

16% of my 5-hour usage limit was used by a simple investigation prompt to examine my claude.md file using opus 4.6 200k context window. No subagents, mcp servers, long context window or anything else can explain the insane consumption. At this rate, I'll be out of tokens the second I actually do something more complex.

What Should Happen?

Prompt should have used <<1% of my usage limit given its simple nature. Just two days ago, I was using Opus 4.6 1M context with planning complex architecture and writing code non-stop for hours and I wasn't hitting my usage limit. Today, I'm just a few prompts deep and my usage is being eaten away for simple prompts with very little context. I don't know what's going on.

Error Messages/Logs

Steps to Reproduce

Use claude opus 4.6 in any sort of rigorous context. It's blowing up usage limits at the moment.

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

2.1.97

Claude Code Version

2.1.104 (Claude Code)

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_