[BUG] Fin AI agent claims it has no ability to pass messages to a human support

Resolved 💬 2 comments Opened May 17, 2026 by marcindulak Closed Jun 16, 2026

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?

Fin AI agent refuses to pass messages to human support, which contradicts its responsibilities described on https://support.claude.com/en/articles/9015913-how-to-get-support, which states that

If your issue requires additional investigation or input from our Product Support team, Fin will pass your inquiry along and someone will respond via email.

<img width="985" height="459" alt="Image" src="https://github.com/user-attachments/assets/5a3e58b9-08b2-4df2-87b8-5de2449862a5" />

<br>

Here is the Fin AI agent response

<img width="391" height="477" alt="Image" src="https://github.com/user-attachments/assets/e20ac002-0dd9-4ada-afb1-cacb31b04209" />

<br>

This is a regression, because in February 2026 the agent passed a message to human support, that resulted in the following email response in March 2026.

<img width="835" height="484" alt="Image" src="https://github.com/user-attachments/assets/6aa68942-20c2-4117-9983-6887ce6c4fa6" />

What Should Happen?

Fin AI agent passes the message to human support.

Error Messages/Logs

Steps to Reproduce

Start "Fin AI agent", describe the early resest of weekly limits problem https://github.com/anthropics/claude-code/issues/49616#issuecomment-4470282130 and require the agent to pass it to human support.

Claude Model

Not sure / Multiple models

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

2.1.116 (Claude Code)

Platform

Anthropic API

Operating System

Other Linux

Terminal/Shell

Other

Additional Information

_No response_

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗