[BUG] Critical: Token consumption is 100x faster after v2.1.88 leak.

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?

After the recent CLI update/leak, Claude Code is re-reading the whole chat history for every prompt. My 5-hour daily limit is finishing in just 21 minutes. This is a massive token drain.

What Should Happen?

Tokens should be cached properly, and usage should be normal as it was in v2.1.87.

Error Messages/Logs

Steps to Reproduce

Open Claude Code: Start a session in any project folder.

Sequential Prompts: Send 5-6 very simple prompts (e.g., "Explain this file," "Add a comment," "Rename a variable").

Observe Bug: Notice that the 5-hour usage limit is exhausted in ~ 21 minutes because the model re-reads the full chat history and file context for every single prompt instead of using cache.

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

claude-opus-4-6[1m]

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

VS Code integrated terminal

Additional Information

_No response_