[BUG] Scheduled task runs incorrectly marked "Skipped" after previous agent completes normally (Claude code desktop)

Resolved 💬 3 comments Opened Apr 12, 2026 by fpalsson Closed Apr 16, 2026

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 a scheduled task fires and the agent completes normally, subsequent scheduled runs are incorrectly marked "Skipped" in the Claude Code desktop GUI, as if the previous agent is still active. The agent had fully exited — no active session existed — yet the concurrency guard continued blocking new runs for multiple hours.

Actual behavior

4 consecutive hourly runs (covering a ~4 hour window) were marked "Skipped" in the desktop GUI. The task had to be manually re-enabled and a new conversation had to be started to clear the false "active session" state.

Environment

Platform: Windows 11 Pro
Claude Code: Desktop app
Schedule: 0 (hourly, with jitter)

What Should Happen?

Schedules should not have been skipped.

Error Messages/Logs

Steps to Reproduce

Create a scheduled task with an hourly cron schedule
The task runs a Python script via Bash and exits cleanly (exit code 0)
The agent completes and the session ends normally
Observe subsequent hourly trigger(s) in the desktop GUI
Expected behavior

Each hourly trigger fires a new agent, since the previous one has completed.

Does not happen every time. Seems to be more affected by manual runs, since user automatically watching the session run.

Claude Model

Sonnet (default)

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.96

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

Other

Additional Information

_No response_

View original on GitHub ↗

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