-
I am following this tutorial on Microsoft Learn: Develop an AI agent with Azure AI Foundry Agent Service The training is using Azure Foundry AI SDK to connect to an Azure Foundry Agent, and send prompt . when I ask the prompt to generate that file, it reply that the generated file is saved in "sandbox:/mnt/data/...." And the training is assuming I am running the code from Azure Cloud Shell. But I am running the code from my machine. |
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 3 replies
-
Hi @gkarwchan The sandbox path is the internal storage used by the agent itself. The app code in the lab should then retrieve that and save it to the same folder as the Python executable - worth taking a look there. Just a note - the cloud shell approach is the recommended mostly because we can't control local setups that might prevent the code from running as expected etc |
Beta Was this translation helpful? Give feedback.
-
Could we edit the name of this post so it useful for other in the future @gkarwchan - maybe something like Running Agent Learn Module on own machine |
Beta Was this translation helpful? Give feedback.
-
Hi @gkarwchan - just a follow up, the lab has been updated as there were changes to the CodeInterpreter tool - there should be a new path through the learning resource now. We really appreciate you highlighting this in the community so we could support! |
Beta Was this translation helpful? Give feedback.
Hi @gkarwchan
The sandbox path is the internal storage used by the agent itself. The app code in the lab should then retrieve that and save it to the same folder as the Python executable - worth taking a look there.
Just a note - the cloud shell approach is the recommended mostly because we can't control local setups that might prevent the code from running as expected etc