Hi team, I need help I want to install boto3 to connect to AWS from my flyte sandbox environment running locally I tried adding the boto3 python package in the ImageSpec in my workflow but when I register the workflow it does not install it inside the container and the workflow fails. Any support for this? below is the code: and the command I used to register the workflow is: pyflyte register workflows/aws_identity.py --project flytesnacks --domain development Code: from flytekit import task, workflow, ImageSpec image_spec = ImageSpec( packages=["boto3", "pandas"], registry="ghcr.io/flyteorg", ) if image_spec.is_container(): import boto3 @task(container_image="image_spec") def whoami() -> str: import boto3 client = boto3.client("sts") identity = client.get_caller_identity() return f"Hello from account: {identity['Account']} as {identity['Arn']}" @workflow def identity_wf() -> str: return whoami()

registry="ghcr.io/flyteorg",

Unless you have access to flyte's GH image registry, the value for

registry

should generally be your private container registry (either locally or in your AWS account)

Thank you for the quick response! Let me try it out

@numerous-pizza-94528 did the above work for you?