It is not clear how MLClient searches for config.json. It seems that from practice, it does not need to be in the current directory of the code and is discoverable if it is at one of the parents of the current directory; but this behavior is not documented.

Also, MLClient's from_config() does not expect config.json to be in ./.azureml directory, which is the default path for workspace.write_config() https://learn.microsoft.com/en-us/python/api/azureml-core/azureml.core.workspace.workspace?view=azure-ml-py#azureml-core-workspace-workspace-write-config . Would be great to be consistent about this, or at least document their behavior in more detail.