Not enough memory to start a new query

[richard-yates]

On long-running sessions, you hit the error 'Not enough memory to start a new query' very easily. This then renders that session inoperable within clagentic-console, however running directly in Claude runs fine

[clagentic-triage]

Thanks for the report. This looks like a session memory management issue in the daemon—definitely in scope.

To help us investigate, could you provide:

1. Version: What version of @clagentic/console are you running? (Check with clagentic-console --version or look at package.json)
2. Reproduction steps: More detail on the scenario:
   • What type of operations trigger the error? (e.g., large file edits, many agent calls, long prompts)
   • Roughly how long is the session running before the error appears?
   • Does it happen consistently or intermittently?
3. Environment:
   • What OS and how much free RAM when the error occurs?
   • How many concurrent sessions are running?

Once we have this, we can look at session memory management in the daemon and understand why the error surfaces in clagentic-console but not in Claude directly.