[BUG] Visible thinking uses whimsical/cutesy verbs even after explicit instruction to stop

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?

In Claude Code (VSCode extension), the model's visible thinking stream repeatedly uses playful/cutesy verbs for routine actions (e.g. "combobulating," "spelunking," "frolicking," "moseying," "scheming") instead of plain ones ("working," "checking," "looking"). This happened repeatedly in a session even after being told directly, multiple times, to stop.

This is used in a professional work environment (enterprise software development), not a casual/consumer context, and the tone reads as unprofessional there.

What Should Happen?

Request: a way to enforce plain, non-whimsical language in the thinking stream (e.g. respecting user/project instructions the same way final output does), or a setting to disable stylized phrasing project-wide.

Error Messages/Logs

Steps to Reproduce

Ask literally anything. as it thinks its is "Booping" , "cooking", "Brewing"...

Claude Model

Sonnet (default)

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.268

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

VS Code integrated terminal

Additional Information

its literally 100% of every question you ask it