Azure.AI.Projects (.NET)
Build AI agents with Azure AI Foundry for .NET
Developers struggle to integrate Azure AI Foundry services into .NET applications. This skill provides complete SDK guidance for agents, connections, datasets, and evaluations with production-ready patterns.
下载技能 ZIP
在 Claude 中上传
前往 设置 → 功能 → 技能 → 上传技能
开启并开始使用
测试它
正在使用“Azure.AI.Projects (.NET)”。 Create a math tutor agent that can solve equations step by step
预期结果:
Agent created with ID 'agent-12345'. Instructions set to guide users through equation solving with step-by-step explanations. Thread initialized and ready for user queries.
正在使用“Azure.AI.Projects (.NET)”。 Evaluate my agent's responses for relevance and accuracy
预期结果:
Evaluation job 'eval-67890' completed. Relevance score: 0.87, Accuracy score: 0.92. 15 of 20 test cases passed. Detailed report available in Azure AI Foundry portal.
安全审计
安全This is a documentation-only skill providing guidance for Azure AI Projects SDK for .NET. Static analysis found 0 security issues with a risk score of 0/100. The skill contains code examples that follow Azure security best practices including DefaultAzureCredential for authentication and async patterns for I/O operations. No executable code, network calls, or suspicious patterns detected.
质量评分
你能构建什么
Enterprise AI Agent Development
Build production AI agents with persistent threads, tool integration, and versioned deployments for enterprise applications.
AI Model Evaluation Pipeline
Create automated evaluation workflows for AI models using built-in evaluators for relevance, fluency, and groundedness metrics.
RAG Solution with Azure AI Search
Implement retrieval-augmented generation solutions connecting AI agents to Azure AI Search indexes for knowledge-based responses.
试试这些提示
Help me create a simple persistent agent using Azure.AI.Projects SDK. I need an agent that can answer customer support questions using the gpt-4o-mini model. Show me the authentication setup and basic agent creation code.
I want to create an agent that can search through uploaded documentation files. Show me how to configure the FileSearchToolDefinition, upload a dataset, and attach it to a persistent agent in Azure AI Foundry.
Explain how to use Azure.AI.Projects.OpenAI package to create versioned agents. I need to deploy an agent with web search tools, get the response client, and manage agent versions for production rollout.
Help me set up an evaluation pipeline for my AI agent. Show me how to configure the Relevance evaluator, create an input dataset, run the evaluation, and interpret the results using the EvaluationsClient.
最佳实践
- Use DefaultAzureCredential for production authentication to leverage managed identities and avoid hardcoded credentials
- Always use async methods for I/O operations and poll with 500ms delays when waiting for agent runs to complete
- Clean up resources by deleting threads, agents, and datasets when no longer needed to avoid unnecessary costs
避免
- Do not use includeCredentials: true unless you specifically need credential values - this exposes sensitive information
- Avoid synchronous methods in production code as they block threads and reduce scalability
- Do not store agent or thread IDs in client-side code without proper validation and access controls