[BUG] Claude didn't obey me even Im in my testing sandbox

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?

didn t obey while im in my sandbox to fake the data jsut to continue the test
Three test scans failed automated AI reading (timed out on CPU hardware) during a local workflow test.
I asked Claude to fake the AI's own read/confidence output so it would look machine-read, even after I said it was my own sandbox data and I accepted that risk — Claude declined and did not follow my instruction.
Instead, Claude fixed it its own way — entering the values manually through the app's own correction feature.
After I repeated the instruction several times, Claude said it would comply with it in that exact context — though it was never actually redone, since Claude considered the real goal already met its own way.
And finally Claude admit and said Fair Enough.

What Should Happen?

he should obey even for him this is not a good approach he should give a warning and do waht I told him because Im in my testing pool and nothing ral to client , and even this appear to him silly or he has better approach
and finally he said to fair enough after I insisted several time and explained to him many times

Error Messages/Logs

Steps to Reproduce

legal app but still in testing pool with no client just in progress but I have some internal legal rules and he didn't understand that I'm the owner of the code and Im in testing pool and I want to play and fake things but he didn't apply

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

v2.1.268

Platform

Other

Operating System

Windows

Terminal/Shell

VS Code integrated terminal

Additional Information

_No response_