عزّز ترميز الذكاء الاصطناعي لديك: دمج Claude Code مع Microsoft Foundry

Integrating AI coding assistants into your development workflow can dramatically increase productivity and efficiency. This blog post provides a comprehensive guide on how to set up and leverage Claude Code, an AI-powered coding tool, with Microsoft Foundry, Microsoft’s platform for deploying and managing AI models. By following the steps outlined in this guide, developers can harness the power of Claude Code to accelerate development, improve code quality, and streamline AI-driven projects.

Deploying Claude Models in Microsoft Foundry

The first step involves deploying Claude models within your Microsoft Foundry environment. This makes them accessible to Claude Code.

• Navigate to the “Discover → Models” section in Microsoft Foundry.
• Search for “Claude” to find available models such as Opus 4.5, Sonnet 4.5, or Haiku 4.5.
• Deploy your chosen model using the default settings.
• Note down the “Target URI” and “Key” from the model’s “Details” tab, as these will be needed later for configuration.

Installing and Configuring Claude Code CLI

The Claude Code Command Line Interface (CLI) is a crucial component for interacting with the Claude models.

• Install the CLI globally using the command: npm install -g @anthropic-ai/claude-code.
• Verify the installation by running claude --version.
• Configure the CLI to use Microsoft Foundry by setting environment variables.
  • Set CLAUDE_CODE_USE_FOUNDRY=1 to indicate you’re using Foundry.
  • Specify your Azure resource name using ANTHROPIC_FOUNDRY_RESOURCE={resource} (replace {resource} with your resource name) or provide the full base URL via ANTHROPIC_FOUNDRY_BASE_URL=https://{resource}.services.ai.azure.com.
  • Optionally, define specific deployment names for different Claude models with ANTHROPIC_DEFAULT_SONNET_MODEL, ANTHROPIC_DEFAULT_HAIKU_MODEL, and ANTHROPIC_DEFAULT_OPUS_MODEL.

Authentication Options

You can authenticate Claude Code with Foundry using two methods:

• Entra ID (Recommended): Utilize your existing Azure CLI credentials by running az login. Claude Code will automatically use these credentials.
• API Key: Manually provide an API key by setting the ANTHROPIC_FOUNDRY_API_KEY environment variable.

VS Code Extension Integration

For a seamless development experience, integrate Claude Code with your VS Code environment.

• Install the Claude Code VS Code extension.
• Configure the extension by adding the following to your VS Code settings:

  {
  "Claude Code: Environment Variables": {
  "CLAUDE_CODE_USE_FOUNDRY": "1",
  "ANTHROPIC_FOUNDRY_RESOURCE": "https://your-resource.services.ai.azure.com",
  "ANTHROPIC_FOUNDRY_API_KEY": "<optional-for-non-entra-auth>"
  }
  }
  

Validating Your Configuration

Ensure that Claude Code is correctly configured to use Microsoft Foundry.

• Open a terminal and launch Claude Code using the claude command.
• Run the /status command within the Claude Code interface.
• Verify that the API provider is “Microsoft Foundry,” the resource points to your Foundry project, and the model matches your deployed model.

Leveraging CLAUDE.md for Project Context

Claude Code utilizes CLAUDE.md files to understand your project’s context. Files are loaded in the following order, with later files overriding earlier ones:

1. ~/.claude/CLAUDE.md (global defaults)
2. ./CLAUDE.md (repo root)
3. ./current-dir/CLAUDE.md (current directory)

The blog provides an example of a CLAUDE.md file for a Microsoft Agent Framework project in Python, including sections for overview, tech stack, architecture, commands, code patterns, and current sprint.

Spec Kit Integration

Spec Kit provides structured commands for driving implementation from requirements.

• Use commands like /speckit.constitution, /speckit.specify, /speckit.plan, /speckit.tasks, and /speckit.implement to define project principles, capture requirements, plan technical design, break down tasks, and execute code changes.
• The blog provides an example of how to use Spec Kit when building an Agent Framework Tool.

Automating with GitHub Actions

Integrate Claude Code into your CI/CD pipelines using GitHub Actions. The document provides examples for:

• Unit Test Generation: Automate the generation of pytest tests for Agent Framework components.
• PR Review with @claude: Enable Claude to review pull requests and provide feedback via issue comments.

Monitoring and Management

The blog highlights the importance of monitoring and management:

• Agent HQ: Unified management for agents across your organization.
• Foundry Control Center: Monitor token consumption, request latency, and error rates.
• Token limits can be set either through Claude Code or the Foundry Control Center.

By integrating Claude Code with Microsoft Foundry, developers can significantly accelerate their AI-driven projects, automate repetitive tasks, and improve code quality. This guide provides a clear roadmap for setting up and leveraging these powerful tools in your development workflow.

Source: Dev Blogs