auto-compaction stopped working = conversation fails when manually compacting

Resolved 💬 3 comments Opened Oct 6, 2025 by bharris52468 Closed Oct 10, 2025

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?

Title: Auto-compaction stopped working - conversation fails when manually compacting

Description:

Claude Code was previously auto-compacting conversations when approaching context limits, but this behavior stopped working
approximately on 2025-10-05.

Current behavior:

  1. Conversation grows too long
  2. System tells user to run /compact
  3. When user runs /compact, it fails because conversation is already too long

Expected behavior:

  • Claude Code should automatically compact conversations before reaching critical context limits
  • Users should never be in a situation where they're told to compact but compaction fails

Impact:
This appears to be a regression - the auto-compaction feature was working for weeks but suddenly stopped.

Environment:

  • Platform: Linux 6.14.0-29-generic
  • Date noticed: ~2025-10-05

What Should Happen?

Expected behavior:

  • Claude Code should automatically compact conversations before reaching critical context limits
  • Users should never be in a situation where they're told to compact but compaction fails

Error Messages/Logs

no error other than can't compact conversation too long

Steps to Reproduce

work until you can't because the conversation gets too long

Claude Model

Not sure / Multiple models

Is this a regression?

Yes, this worked in a previous version

Last Working Version

worked yesterday and didn't update versions that I'm aware of

Claude Code Version

2.0.8

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

Other

Additional Information

_No response_

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗