I’m seeing inconsistent behavior when DebugMCP handles long-running operations during step_over.

Scenario 1 – Function with sleep

demo_run.py

from utils import demo
def main() -> None:
 print('Start')
 demo()
 print('End')
if __name__ == "__main__":
 main()

utils.py

from time import sleep
def demo():
 sleep(100)
 print('in demo')

Failure Behavior:

When execution reaches demo() and sleep(100) is still running, step_over fails with:
Debug session is not ready. Please wait for initialization to complete.
Attempting stop_debugging afterward fails with:
No active debug session to stop

In the end the debug session is still running and must be terminated manually.

Scenario 2 – Sleep inside main

from time import sleep
def main() -> None:
 print('Start')
 sleep(100)
 print('End')
if __name__ == "__main__":
 main()

Result: step_over times out while sleep(100) is still executing, and the agent later fails to stop debugging.

Issues:

1. step_over times out or reports invalid session while code is still executing.
2. stop_debugging fails to terminate an active session after such errors.

Expected Behavior:

• Long-running execution should not invalidate the session.
• step_over should wait or return a clear "execution in progress" state.
• stop_debugging should reliably stop any active session.