Retrieving downloaded file from GPTs in their runner instance

Hi there,

I want to create a GPT that can clone a repo and store it as knowledge for the chat. So the user can talk with the repo. I created an API that clones a repo.

Where is the best place to clone the repo to in the runner so that custom GPT can have access to it for knowledge? (example: ‘…/mnt/data’ )

I was downloading it to a relative file ‘…/mnt/data’ but when the code interpreter execute code it says the file cannot be found. Does this mean that code interpreter is running in a new instance.

What do you use in your Action spec to ensure the GPT downloads the file? Have you run into any size limits?