Runaway Claude Code session consumed weekly limits during stalled/resumed failure state

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?

Claude Code entered an apparent stalled/resumed failure state during recent service instability/incidents.

The session continued running autonomously for approximately 16 hours while consuming nearly the entire available usage limit without meaningful execution or usable output.

Observed behavior included:

• stalled execution
• resumed-session behavior
• no productive completion
• autonomous limit consumption
• inability to continue using Claude Code afterward due to exhausted limits

The issue appears related to the recent Claude Code / model incidents acknowledged by Anthropic support.

Environment:

• macOS
• Claude Code desktop app
• large active repository/context

This behavior created significant unintended usage consumption during what appeared to be a malfunction state.

What Should Happen?

Claude Code should stop execution when a session becomes stalled or non-responsive instead of continuing autonomous token consumption for hours without meaningful output.

The system should also preserve enough remaining usage to allow bug reporting or feedback submission after a malfunction occurs.

Error Messages/Logs

Steps to Reproduce

1. Open Claude Code on a large active repository/project.

1. Resume or continue a long-running session with substantial context and multiple pending implementation tasks.

1. Allow Claude Code to continue autonomous multi-step execution (editing files, running commands, reading files, creating components, committing waves, etc.).

1. During service instability or degraded model behavior, observe that the session may continue consuming usage autonomously for an extended period.

1. Eventually the session exhausts the available weekly/session limits and becomes unusable, while preventing further reporting through /feedback or continued interaction.

Observed result:

• excessive autonomous usage consumption
• stalled/non-responsive behavior
• abrupt usage exhaustion
• inability to continue or report from inside Claude Code afterward

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

Unknown

Platform

Other

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_