Vouch request: MatejKosec

[MatejKosec]

What do you want to work on?

I am working on running claude inside OpenShell pointed at a custom inference provider.

Why this change?

I noticed that when using a custom provider the token_counts part of the API is blocked. Claude Code uses this API to determine whether to compress context before sending the next request or not and to provide an accurate token count in the detailed output.

Checklist

• I wrote this request myself (not AI-generated)
• I have read CONTRIBUTING.md