What problem does it solve?
This Skill reduces friction when building and deploying container-based hosted agents into Azure AI Projects by clarifying required SDK imports, authentication patterns, ACR prerequisites, and agent definition parameters so engineers can reliably create and manage custom agent versions.
Core Features & Use Cases
- Hosted agent provisioning: Guidance to create ImageBasedHostedAgentDefinition instances and call create_version to register container-based agents.
- Resource, tools, and environment configuration: Instructions for specifying cpu, memory, tools, and environment_variables for container runtime.
- Lifecycle operations and troubleshooting: How to list and delete agent versions and address common errors such as ImagePullBackOff, missing permissions, and protocol configuration mismatches.
- Use Case: Deploy a data-processing agent that runs custom code in a container image stored in Azure Container Registry, integrates a code interpreter tool, and communicates with an Azure AI Projects endpoint.
Quick Start
Create a hosted agent by providing your ACR image, granting AcrPull to the project's managed identity, and invoking AIProjectClient.create_version with DefaultAzureCredential.