Issue Station
Sign In
Tools hereToolbox

claude-code - 💡(How to fix) Fix [BUG] Claude Code fully consumes the usage limits with a single task that uses only about 10% of the context.

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…

Error Message

Error Messages/Logs

RAW_BUFFERClick to expand / collapse

Preflight Checklist

  • I have searched existing issues and this hasn't been reported yet
  • This is a single bug report (please file separate reports for different bugs)
  • I am using the latest version of Claude Code

What's Wrong?

Since this morning, I’ve been experiencing a complete consumption of my 5-hour usage limits in Claude Code from a single prompt. This situation is frankly absurd. It happens without any apparent reason, and I’ve never encountered anything like this before.

I waited three hours and tried again. I launched a prompt using Claude Code (Mac, 4.7 High), and within just five minutes, the entire session limit was consumed—despite having no active tasks running.

<img width="883" height="453" alt="Image" src="https://github.com/user-attachments/assets/4352ee8a-1cde-49b3-847e-c27bf99a45c6" />

What Should Happen?

After completing the task, I noticed that the limits continue to be consumed even though there are no active processes or background tasks running. This has already happened twice, exhausting all the session limits.

Error Messages/Logs

Steps to Reproduce

no step to reproduce. only i'm using plugin claude-mem but i'm using it'f from month and never see this situation

Claude Model

Opus 4,7

Is this a regression?

I don't know

Last Working Version

No response

Claude Code Version

last

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

Powershell

Additional Information

No response

extent analysis

TL;DR

The issue might be related to the claude-mem plugin, and disabling or investigating its usage could help mitigate the unexpected consumption of session limits.

Guidance

  • Investigate the claude-mem plugin configuration and usage to see if there are any recent changes or updates that could be causing the issue.
  • Try disabling the claude-mem plugin temporarily to see if the session limit consumption stops.
  • Review the Anthropic API documentation to understand how session limits are tracked and consumed, as this might provide insight into what is causing the limits to be reached so quickly.
  • Consider reaching out to the plugin author or the Anthropic API support team for further assistance, as the issue may be related to the interaction between the plugin and the API.

Notes

The lack of error messages or logs makes it difficult to pinpoint the exact cause of the issue. The fact that the claude-mem plugin has been in use for a month without issues suggests that there may be a recent change or update that is causing the problem.

Recommendation

Apply workaround: Disable the claude-mem plugin temporarily to see if the issue resolves, as this is the most likely cause based on the provided information.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#agent setup#task chaining#parallel task#integration issue

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

claude-code - 💡(How to fix) Fix [BUG] Claude Code fully consumes the usage limits with a single task that uses only about 10% of the context.