Google ai studio code execution

Google AI Studio
1

in google ai studio, why does code execution sometimes show google:python_interpreter and ask enter function response instead of the executed results? It seems like a bug or a limitation being triggered

by argon:

1. Not a Python CPU timeout (the simulation is super fast).
2. It is a platform/orchestrator glitch where the system failed to execute the tool call and return control back to the model to finish writing the message.
2

No reply after 2 days?
This bug is still present on ai studio. In the current case, when it runs 11 times of correct code execution, the 12th time becomes google:python_interpreter instead of getting real execution results. It feels like a Jupyter Notebook with only limited cells; above the cell limit, the cells can no longer run.
This is annoying, fix it please.