azure-ai-projects-py

What This Skill Does

The azure-ai-projects-py skill provides a powerful interface for interacting with the Azure AI Foundry platform using the official Python SDK. It serves as your primary bridge to managing the end-to-end lifecycle of AI applications in an enterprise Azure environment. This skill allows you to programmatically manage project assets, including model deployments, connections to external services, dataset management, and indexing for RAG pipelines. It is the high-level management layer that sits above your specific agent implementations.

Installation

To integrate this skill into your environment, use the OpenClaw CLI:

clawhub install openclaw/skills/skills/thegovind/azure-ai-projects-py

You will also need the relevant Python dependencies installed in your execution environment:

pip install azure-ai-projects azure-identity

Ensure your AZURE_AI_PROJECT_ENDPOINT is correctly set in your environment variables to authorize the SDK to communicate with your specific Azure AI Project.

Use Cases

• Project Orchestration: Automate the creation and configuration of AI Foundry projects, including setting up connections to data stores or model endpoints.
• Agent Lifecycle Management: Use the AIProjectClient to programmatically create, update, and version control your agents, including those defined via PromptAgentDefinition.
• Evaluation Pipelines: Run automated evaluations against your models to track performance metrics and drift over time.
• RAG Implementation: Manage index creation and dataset uploads, facilitating easier RAG workflows for your AI agents.
• Unified API Usage: Seamlessly switch between Foundry-specific management operations and standard OpenAI-compatible API calls using the get_openai_client() method.

Example Prompts

1. "Initialize the Azure AI Project client using my environment variables and list all available model deployments in the project."
2. "Create a new versioned agent using PromptAgentDefinition with the gpt-4o-mini model and standard instructions for a technical support bot."
3. "Fetch the current list of datasets from my AI project and run a simple evaluation on the 'test-set-01' using the evaluation module."

Tips & Limitations

• Hierarchy: Remember that this is a project-management SDK. If you need low-level, individual agent threading and run control, you should pair this with the azure-ai-agents-python skill.
• Authentication: Always utilize DefaultAzureCredential to ensure your local development flows match your production identity management seamlessly.
• Versioning: When working with agents, lean on the PromptAgentDefinition for consistent and reproducible deployments in production pipelines.
• Scope: This skill requires an active Azure AI Project; it cannot perform operations on generic OpenAI endpoints without a valid Azure Foundry project context.