[BUG] Reaching the 5-hour session limit kills in-flight workflow agents instead of pausing them
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?
When the 5-hour session limit is reached in the middle of a Workflow run, the agents still running fail with "You've hit your session limit" instead of pausing. On 2026-09-28, 26 in-flight agents of one run failed this way. Plain background agents (the Agent tool) have no pause at the limit either.
Agents that had already finished stay saved, and resumeFromRunId after the reset re-runs only the failed ones, so no finished work is lost, but the run has to be resumed by hand.
What Should Happen?
The 2.1.271 changelog says dynamic workflows pause at the usage limit and continue automatically after the reset instead of dropping agents. The run should pause and continue on its own after the reset.
Error Messages/Logs
Steps to Reproduce
- Start a Workflow that runs many agents in parallel (here a research run of 26+ agents) close to the 5-hour session limit.
- Let the limit be reached while agents are still running.
- The running agents end with "You've hit your session limit"; the run neither pauses nor continues after the reset.
- After the reset, re-run with
resumeFromRunId: only the failed agents run again.
Claude Model
_No response_
Is this a regression?
No, this never worked
Last Working Version
_No response_
Claude Code Version
2.1.281 (Claude Code)
Platform
Anthropic API
Operating System
Windows
Terminal/Shell
Other
Additional Information
Claude desktop app (Code tab), Windows 11 Home 10.0.26200. Not re-run on a later version, because reproducing it means exhausting the 5-hour limit on purpose; no changelog entry through 2.1.285 mentions a fix.
This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗