mcpField
M
microsoft-foundry-classic

microsoft-foundry-classic

github.com/microsoftdocsAgent 工作流796181 次安装3 天前更新
代码仓库

关于 microsoft-foundry-classic

microsoft-foundry-classic 是面向使用 Microsoft Foundry Classic 的 Agent 的指导技能，结合本地参考资料与远程文档获取，支持模型、Agent、应用和部署决策。

主要功能

  1. 01故障排查与配置指导
  2. 02模型选择、部署和成本规划
  3. 03安全、配额与架构参考
  4. 04本地参考与远程文档获取

适用场景

  1. 01诊断部署、监控或网络问题
  2. 02选择模型、区域和部署选项
  3. 03规划安全的 Foundry Agent 与应用

MCP 相关

其文档索引涵盖 MCP 集成，并建议使用 mcp_microsoftdocs:microsoft_docs_fetch 获取远程文档，以 fetch_webpage 作为备用方式。

安装

复制下面的命令，在终端中运行。

$npx skills add https://github.com/microsoftdocs/agent-skills --skill microsoft-foundry-classic

安装后在 Claude Code 中输入/microsoft-foundry-classic确认已加载。

详细介绍

Microsoft Foundry Classic 技能

此技能提供有关 Microsoft Foundry Classic 的专业指导，涵盖故障排除、最佳实践、决策、架构与设计模式、限制与配额、安全性、配置、集成与编码模式，以及部署。它结合了本地快速参考内容与远程文档获取功能。

如何使用此技能

IMPORTANT for Agent：使用下方的类别索引查找相关部分。对于带有行范围的类别（例如，L35-L120），使用 read_file 指定相应行。对于带有文件链接的类别（例如，[security.md](security.md)），使用 read_file 读取链接的参考文件

IMPORTANT for Agent：如果 metadata.generated_at 的日期早于 3 个月以上，建议用户从代码库拉取最新版本。如果 mcp_microsoftdocs 工具不可用，建议用户安装该工具：安装指南

此技能需要网络访问权限才能获取文档内容：

  • 首选：使用 mcp_microsoftdocs:microsoft_docs_fetch，并传入查询字符串 from=learn-agent-skill。返回 Markdown。
  • 备用：使用 fetch_webpage，并传入查询字符串 from=learn-agent-skill&accept=text/markdown。返回 Markdown。

类别索引

类别行描述
故障排除L37-L46诊断和修复 Foundry 问题：Prompt Flow 计算会话、部署/监视、专用终结点、Azure OpenAI 微调、风险与安全警报，以及已知的门户问题。
最佳实践L47-L63设计安全、高效的 Azure OpenAI/Foundry 应用：系统消息、安全模板、微调（包括视觉/推理）、延迟/吞吐量、令牌预算，以及聊天应用评估。
决策L64-L91有关在 Foundry 中选择模型、区域、部署和计费方式，以及迁移路径的指导，包括 PTU 规模估算、生命周期/退役规划和灾难恢复决策。
架构与设计模式L92-L98设计多 Agent 架构、配置 Foundry Agent Service 以实现复原能力，并了解模型路由器的行为、故障转移和灾难恢复策略。
限制与配额L99-L114Foundry Agent、模型（包括 Claude、Azure OpenAI）的配额、速率限制和区域可用性，以及如何管理、申请和提高部署与吞吐量配额。
安全性L115-L160Foundry 的安全、隐私和合规性：RBAC、Entra ID、网络隔离、CMK、Azure Policy、护栏/内容筛选器，以及安全地使用模型、工具和“基于你的数据”功能。
配置L161-L210配置、监视和排查 Foundry classic：Agent、评估器、模型、存储、网络、计算、OpenAI 集成、可观测性和自动化（Terraform、SDK）。
集成与编码模式L211-L307将 Foundry classic Agent 和模型与外部工具、数据和 API 集成（Azure AI Search、Bing、SharePoint、MCP、OpenAPI、LangChain、SK），以及 Azure OpenAI 的使用、微调、RAG 和实时音频。
部署L308-L327大规模部署 Foundry 中心/模型：区域、部署类型、Bicep/Terraform/CLI、托管和无服务器计算、HF 模型、微调模型、评估，以及区域功能支持
故障排查
主题URL
排查 Foundry 中 Prompt Flow 计算会话问题https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/prompt-flow-troubleshoot
排查 Foundry 部署和监视问题https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/troubleshoot-deploy-and-monitor
排查 Foundry 专用终结点连接错误https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/troubleshoot-secure-connection-project
诊断并修复 Azure OpenAI 微调问题https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-troubleshoot
在 Foundry 中监视并排查风险与安全问题https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/risks-safety-monitor
解决 Microsoft Foundry classic 门户中的已知问题https://learn.microsoft.com/en-us/azure/foundry-classic/reference/foundry-known-issues
最佳实践
主题URL
为 Azure OpenAI 设计有效的系统消息https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/advanced-prompt-engineering
在 Azure OpenAI 中应用安全系统消息模板https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/safety-system-message-templates
为 Azure OpenAI 编写安全系统消息https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/system-message
在 Foundry 中管理聊天补全和令牌预算https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/chatgpt
对 Azure OpenAI 模型使用直接偏好优化https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-direct-preference-optimization
在 Foundry 模型微调时应用安全评估https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-safety-evaluation
在 Foundry 中使用图像微调 GPT-4 视觉模型https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-vision
在 Foundry 中优化 Azure OpenAI 延迟和吞吐量https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/latency
应用 Azure OpenAI On Your Data 的最佳实践https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/on-your-data-best-practices
优化 Azure OpenAI 预测输出以降低延迟https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/predicted-outputs
在生产环境中运行 Foundry 预配吞吐量https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/provisioned-get-started
对 Foundry 中的推理模型应用强化微调https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/reinforcement-fine-tuning
使用 SDK 评估和改进基于 Foundry 的聊天应用https://learn.microsoft.com/en-us/azure/foundry-classic/tutorials/copilot-sdk-evaluate
决策
主题URL
为 Foundry Agents 选择 Azure OpenAI 模型和区域https://learn.microsoft.com/en-us/azure/foundry-classic/agents/concepts/model-region-support
决定何时以及如何在 Foundry 中微调模型https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/fine-tuning-overview
使用基准测试和排行榜比较 Foundry 模型https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/model-benchmarks
管理 Foundry 托管计算模型的生命周期https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/model-retirement-managed-compute
为 Foundry 选择合适的 Azure 资源类型https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/resource-types
选择 Microsoft Foundry 模型部署类型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/deployment-types
按功能和区域选择 Azure 销售的 Foundry 模型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/models-sold-directly-by-azure
在 Foundry 中选择 GPT-5 还是 GPT-4.1https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/model-choice-guide
将应用程序从 GitHub Models 迁移到 Foundryhttps://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/quickstart-github-models
为 Foundry Agent Service 规划灾难恢复https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/agent-service-disaster-recovery
使用 Foundry 排行榜比较模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/benchmark-model-in-catalog
规划和管理 Microsoft Foundry 中心的成本https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/costs-plan-manage
选择和配置 Microsoft Foundry SDK 终结点https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/sdk-overview
将基于中心的项目迁移到 Foundry classic 项目https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/migrate-project
决定并规划从 Prompt Flow 迁移到 Agent Frameworkhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/prompt-flow-migration-overview
为 Azure Government 部署选择 Foundry 模型https://learn.microsoft.com/en-us/azure/foundry-classic/openai/azure-government
使用 Foundry 模型退役计划规划迁移https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/model-retirement-schedule
根据 Foundry 模型生命周期和退役安排进行规划https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/model-retirements
评估 2024 年 Azure OpenAI 预配吞吐量更新https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/provisioned-migration
为 Foundry PTU 部署选择计费选项https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/provisioned-throughput-billing
估算和管理 Azure OpenAI 微调成本https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-cost-management
为 Foundry 工作负载确定预配吞吐量单位的规模https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/provisioned-throughput-sizing
从预览版迁移到 GA Realtime API 协议https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/realtime-audio-preview-api-migration-guide
决定从存储的 completions 迁移到 Responses API 的路径https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/stored-completions-retirement
架构与设计模式
主题URL
使用 Foundry 连接的 Agent 设计多 Agent 系统https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/connected-agents
从资源和数据丢失中恢复 Foundry Agent Servicehttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/agent-service-operator-disaster-recovery
从平台中断中恢复 Foundry Agent Servicehttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/agent-service-platform-disaster-recovery
限制与配额
主题URL
查看 Foundry Agent Service 的配额和限制https://learn.microsoft.com/en-us/azure/foundry-classic/agents/quotas-limits
评估 Foundry classic 的速率限制和区域支持https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-regions-limits-virtual-network
比较 Foundry 中 Claude 模型的功能、配额和区域https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/claude-models
了解 Foundry 中 Claude 模型的配额和速率限制https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/claude-models-quotas-limits
查阅 Foundry Models 的配额和限制https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/quotas-limits
管理 Microsoft Foundry 中心资源的配额并申请配额https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/hub-quota
管理并提高 Foundry 中的模型部署配额https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/quota
了解 Foundry 预配吞吐量限制和配额https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/provisioned-throughput
使用 Azure OpenAI 全局批处理配额https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/batch
为 Azure OpenAI 部署使用动态配额https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/dynamic-quota
在 Foundry 中管理 Azure OpenAI 配额和速率限制https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/quota
查看 Foundry 中 Azure OpenAI 的配额和限制https://learn.microsoft.com/en-us/azure/foundry-classic/openai/quotas-limits
安全性
主题URL
为 Foundry Agent 安全配置浏览器自动化工具https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/browser-automation
在 Agent 中安全使用 Foundry Computer Use 工具https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/computer-use
使用 Entra ID 和 RBAC 安全访问 Microsoft Foundryhttps://learn.microsoft.com/en-us/azure/foundry-classic/concepts/authentication-authorization-foundry
为 Microsoft Foundry（经典版）配置客户管理的密钥https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/encryption-keys-portal
为 Foundry 中心项目配置客户管理的密钥https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/hub-encryption-keys-portal
为 Microsoft Foundry 中心和项目配置 RBAChttps://learn.microsoft.com/en-us/azure/foundry-classic/concepts/hub-rbac-foundry
为 Azure 销售的 Foundry 模型配置防护措施https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/model-catalog-content-safety
为 Microsoft Foundry 项目配置 RBAC 角色https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/rbac-foundry
了解 Microsoft Foundry 映像的漏洞管理https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/vulnerability-management
为 Foundry 模型配置默认安全和防护策略https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/default-safety-policies
安全配置和管理 Foundry 内容筛选器https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/configure-content-filters
创建自定义 Azure Policy，以限制 Foundry 模型部署https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/configure-deployment-policies
为 Foundry Models 配置 Microsoft Entra ID 无密钥身份验证https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/configure-entra-id
允许 Foundry 托管网络访问本地资源https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/access-on-premises-resources
应用 Azure Policy 来管理 Foundry 中心和项目https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/azure-policy
使用 Azure Policy 控制 Foundry 模型部署https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/built-in-policy-model-deployment
了解 Foundry 模型中的数据处理和隐私https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/concept-data-privacy
为 Foundry 中心配置托管网络隔离https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/configure-managed-network
为安全的 Foundry 项目配置专用链接https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/configure-private-link
安全创建和管理 Microsoft Foundry 中心https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-azure-ai-resource
使用托管 VNet 创建安全的 Microsoft Foundry 中心https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-secure-ai-hub
禁用对 Foundry 中心存储的共享密钥访问https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/disable-local-auth
为 Foundry 评估配置存储帐户和访问权限https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/evaluations-storage-account
为 Microsoft Foundry 中心配置专用链接https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/hub-configure-private-link
为 Foundry 项目配置托管虚拟网络https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/managed-virtual-network
使用 Azure Policy 管理 Foundry 模型部署https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/model-deployment-policy
安全配置 Foundry 操场中的数据聊天https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/secure-data-playground
为 Foundry 设置 Azure Key Vault 连接https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/set-up-key-vault-connection
使用 Content Credentials 处理 Azure OpenAI 图像https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-credentials
在 Foundry OpenAI 中配置和使用个人身份信息筛选https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-filter-personal-information
了解 Azure OpenAI 中的默认 Guardrail 安全策略https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/default-safety-policies
配置和自定义 Microsoft Foundry 内容筛选器https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/content-filters
为 Foundry 中的 Azure OpenAI 配置 Entra ID 身份验证https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/managed-identity
使用虚拟网络和专用终结点保护 Azure OpenAIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/network
配置 Azure OpenAI 网络安全边界https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/network-security-perimeter
为 Azure OpenAI On Your Data 配置网络和访问权限https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/on-your-data-configuration
使用 RBAC 角色管理 Azure OpenAI 访问权限https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/role-based-access-control
为 Azure OpenAI 内容筛选配置自定义阻止列表https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/use-blocklists
为 Foundry 中的 Claude 配置数据隐私和安全性https://learn.microsoft.com/en-us/azure/foundry-classic/responsible-ai/claude-models/data-privacy
为 Foundry 中的 Azure OpenAI 应用版权承诺缓解措施https://learn.microsoft.com/en-us/azure/foundry-classic/responsible-ai/openai/customer-copyright-commitment
了解 Foundry Models 的数据隐私和安全性https://learn.microsoft.com/en-us/azure/foundry-classic/responsible-ai/openai/data-privacy
为 Foundry 中的 Azure OpenAI 应用有限访问策略https://learn.microsoft.com/en-us/azure/foundry-classic/responsible-ai/openai/limited-access
配置
主题URL
配置和排查 Foundry 功能主机https://learn.microsoft.com/en-us/azure/foundry-classic/agents/concepts/capability-hosts
为 Foundry Agent Service 配置标准 Agent 资源https://learn.microsoft.com/en-us/azure/foundry-classic/agents/concepts/standard-agent-setup
使用 Azure Monitor 和 KQL 监视 Foundry Agent Servicehttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/metrics
为 Foundry Agent 连接和配置自己的 Azure 资源https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/use-your-own-resources
为 Foundry Agent Service 配置专用网络https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/virtual-networks
Foundry Agent Service 监视数据参考https://learn.microsoft.com/en-us/azure/foundry-classic/agents/reference/monitor-service
配置和使用内置 Foundry 评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/built-in-evaluators
为 Azure AI Agent 配置 Agent 评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/agent-evaluators
使用 Azure OpenAI 评分器进行模型评估https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/azure-openai-graders
在 Foundry 中创建和配置自定义评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/custom-evaluators
为生成式 AI 配置通用评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/general-purpose-evaluators
配置用于相关性和依据性的 RAG 评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/rag-evaluators
在 Foundry 中配置风险和安全评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/risk-safety-evaluators
在 Foundry 中使用文本相似度评估器https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/evaluation-evaluators/textual-similarity-evaluators
配置和使用 Microsoft Foundry Models 的终结点https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/endpoints
在 AI 项目中配置 Foundry Models 连接https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/configure-project-connection
为 Foundry 模型部署配置 Azure Monitorhttps://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/monitor-models
升级 AI 项目以使用 Foundry 模型部署https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/quickstart-ai-project
选择 Foundry Models 的 SDK 和语言https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/supported-languages
使用自己的 Azure Storage 配置 Microsoft Foundry classichttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/bring-your-own-azure-storage-foundry
为 Foundry Speech/Language 配置客户管理的存储https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/bring-your-own-azure-storage-speech-language-services
配置和管理 Foundry 计算实例https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-manage-compute
在 Foundry classic 中配置和管理 prompt flow 计算会话https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-manage-compute-session
使用 Terraform 预配 Microsoft Foundry (classic)https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-resource-terraform
在基于 Foundry hub 的项目中添加和管理数据https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/data-add
配置对无服务器模型 API 的跨项目访问https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/deploy-models-serverless-connect
通过 Azure ML SDK 和 CLI 创建 Foundry hubhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/create-hub-project-sdk
为 Foundry classic 开发配置工具和 SDKhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/install-cli-sdk
为 Foundry 评估生成合成和模拟数据https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/simulator-interaction-data
使用 OpenTelemetry 跟踪和观测 Foundry Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/trace-agents-sdk
为 prompt flow 部署启用跟踪和反馈https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/trace-production-sdk
使用 Azure 标记隐藏 Foundry 预览功能https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/disable-preview-features
在 Foundry 中创建和使用用于 RAG 的向量索引https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/index-add
为 Foundry AI 应用配置持续监视https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/monitor-applications
监视 prompt flow 应用的质量和令牌使用情况https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/monitor-quality-safety
在 Foundry 中配置和使用 Azure OpenAI v1 APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/api-version-lifecycle
在 Azure OpenAI 中配置内容流式传输和筛选模式https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-streaming
自动化 Azure OpenAI 部署和 TPM 配额https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/automate-quota-deployments
为 OpenAI Batch 配置 Azure Blob Storagehttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/batch-blob-storage
配置和运行 Azure OpenAI 模型评估https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/evaluations
为 Azure OpenAI 配置监视和日志记录https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/monitor-openai
在 Foundry 中为 Azure OpenAI 配置提示词缓存https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/prompt-caching
为 Azure OpenAI 聊天配置可复现输出https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/reproducible-output
为 Foundry 预配部署配置溢出流量https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/spillover-traffic-management
在 Foundry 中配置存储的补全内容和蒸馏https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/stored-completions
Azure OpenAI 监视指标和日志参考https://learn.microsoft.com/en-us/azure/foundry-classic/openai/monitor-openai-reference
集成与编码模式
主题URL
将 Foundry Agent 连接到现有 Azure AI Search 索引https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/azure-ai-search
将 Azure AI Search 集成到 Foundry classic Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/azure-ai-search-samples
将 Azure Functions 集成为 Foundry Agent Service 的工具https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/azure-functions
将 Azure Functions 和 Storage Queues 与 Foundry Agent 配合使用https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/azure-functions-samples
使用 Bing Search grounding 工具和代码示例https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/bing-code-samples
为 Agent 配置 Custom Bing Search groundinghttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/bing-custom-search
在 Foundry classic Agent 中使用 Bing Custom Search groundinghttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/bing-custom-search-samples
使用 Bing Search 结果为 Foundry Agent 提供 groundinghttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/bing-grounding
使用 Foundry Browser Automation 工具自动执行网站任务https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/browser-automation-samples
使用 Foundry Code Interpreter 工具运行 Python 代码https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/code-interpreter
将 Computer Use 工具与 Azure AI Projects SDK 配合使用https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/computer-use-samples
在 Foundry Agent 中维护 classic deep research 工具https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/deep-research-samples
将 Foundry classic Agent 与 Microsoft Fabric 数据集成https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/fabric
使用 Azure AI Agents classic file search 工具https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/file-search
通过代码将文件上传到 Foundry file searchhttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/file-search-upload-files
使用 Foundry Agent Service 实现函数调用https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/function-calling
将 Foundry Agent 连接到外部 MCP 服务https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/model-context-protocol
Foundry Agent 的 MCP 服务集成代码示例https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/model-context-protocol-samples
为 Foundry Agent 配置 OpenAPI 工具和身份验证https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/openapi-spec
在代码中将基于 OpenAPI 的工具与 Foundry Agent 配合使用https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/openapi-spec-samples
使用 SharePoint 工具为 Azure AI Agents 提供 groundinghttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/sharepoint-samples
通过 Azure Logic Apps 事件触发 Foundry classic Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/triggers
Foundry Models 的无服务器 API 推理示例https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/models-inference-examples
调用 Foundry Responses API 生成文本https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/generate-responses
在 Foundry 中部署并调用 Claude 模型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/use-foundry-models-claude
在 Foundry 中部署并调用 FLUX 图像模型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/use-foundry-models-flux
通过 Foundry API 部署并调用 Grok 4.7https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/use-foundry-models-grok
使用 Foundry 聊天模型实现结构化输出https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/use-structured-outputs
在 Foundry 中部署并调用 DeepSeek 推理模型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/tutorials/get-started-deepseek-r1
在 Foundry classic 中配置项目连接https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/connections-add
使用 Azure AI Evaluation SDK 运行本地评估https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/evaluate-sdk
将 LangChain 与 Microsoft Foundry 模型集成https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/langchain
将 LangGraph 和 LangChain 应用连接到 Foundry Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/langchain-agents
在 Microsoft Foundry 部署中使用 LlamaIndexhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/llama-index
使用 Foundry SDK 在云端运行 AI Red Teaming Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/run-ai-red-teaming-cloud
使用 Azure AI SDK 在本地运行 AI Red Teaming Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/run-scans-ai-red-teaming-agent
将 Semantic Kernel 与 Foundry 模型目录配合使用https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/semantic-kernel
在 Foundry 中使用 OpenTelemetry 查看 OpenAI SDK 跟踪记录https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/trace-application
在 VS Code 中为 Foundry Agent 配置 MCP 服务工具https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/develop/vs-code-agents-mcp
部署并使用 CXRReportGen 医疗 AI 模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/healthcare-ai/deploy-cxrreportgen
部署并调用 MedImageInsight 嵌入模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/healthcare-ai/deploy-medimageinsight
使用 MedImageParse 医学图像分割模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/healthcare-ai/deploy-medimageparse
使用 Agent Framework 重建 Prompt Flow 工作流https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/how-to-migrate-prompt-flow-to-agent-framework
在 Foundry 中创建和管理中心范围的连接https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/hub-connections-add
将 Azure AI Inference SDK 应用迁移到 OpenAI SDKhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/model-inference-to-openai-migration
解读 Foundry 中 Azure OpenAI 的 Guardrail 注释https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-filter-annotations
使用文档嵌入为 Foundry Guardrails 格式化提示词https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-filter-document-embedding
在 Foundry 中使用 RAG 的 groundedness 检测https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-filter-groundedness
在 Foundry 中对大模型输出应用受保护内容检测https://learn.microsoft.com/en-us/azure/foundry-classic/openai/concepts/content-filter-protected-material
在 Azure OpenAI 中实现 Computer Use Agenthttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/computer-use
使用 Azure OpenAI 图像生成和编辑 APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/dall-e
使用 gpt-6-luna 调用 Azure OpenAI Decisions APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/decisions
使用 Azure OpenAI Responses API 进行深度研究https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/deep-research
将 Azure.AI.OpenAI .NET 1.x beta 迁移到 2.xhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/dotnet-migration
在 Foundry 中使用 Azure OpenAI 生成嵌入https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/embeddings
使用 Chat Completions 和 Evaluations 测试微调模型https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tune-test
通过 Python、REST 或门户微调 Foundry 模型https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning
在 Foundry 中微调 Azure OpenAI 函数调用https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/fine-tuning-functions
在 Foundry 中使用 Azure OpenAI 实现函数调用https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/function-calling
通过 API 使用支持视觉的 Azure OpenAI 聊天模型https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/gpt-with-vision
配置 Azure OpenAI JSON mode 响应https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/json-mode
将 Azure OpenAI 应用迁移到 OpenAI Python v1.xhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/migration
将 Azure OpenAI 迁移到 JavaScript SDK v4https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/migration-javascript
在 Azure 中通过 SIP 集成 GPT Realtime 音频https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/realtime-audio-sip
在 Azure OpenAI 中通过 WebRTC 实现 GPT Realtime 音频https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/realtime-audio-webrtc
通过 WebSockets 使用 Azure OpenAI GPT Realtimehttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/realtime-audio-websockets
使用工具和流式处理实现 Azure OpenAI Responses APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/responses
在 Azure OpenAI 模型中使用结构化输出https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/structured-outputs
在 OpenAI 和 Azure OpenAI 之间切换 Python 代码https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/switching-endpoints
配置 Azure OpenAI API 事件 Webhookhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/webhooks
将 Azure OpenAI 微调与 Weights & Biases 集成https://learn.microsoft.com/en-us/azure/foundry-classic/openai/how-to/weights-and-biases-integration
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry Models 中使用 Azure OpenAI v1 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/latest
在 Foundry 中使用 Realtime 和 Voice Live 音频事件https://learn.microsoft.com/en-us/azure/foundry-classic/openai/realtime-audio-reference-ga
调用 Azure OpenAI 图像和音频 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/reference
集成 Azure OpenAI 预览版 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/reference-preview
使用 Azure OpenAI v1 预览版 REST APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/reference-preview-latest
使用 API 将 Azure OpenAI 与 Azure Search 数据集成https://learn.microsoft.com/en-us/azure/foundry-classic/openai/references/azure-search
通过 API 将 Azure OpenAI 与 Azure Cosmos DB 数据集成https://learn.microsoft.com/en-us/azure/foundry-classic/openai/references/cosmos-db
通过 API 将 Azure OpenAI 与 Elasticsearch 数据集成https://learn.microsoft.com/en-us/azure/foundry-classic/openai/references/elasticsearch
通过 API 将 Azure OpenAI 与 MongoDB Atlas 配合使用https://learn.microsoft.com/en-us/azure/foundry-classic/openai/references/mongo-db
通过 Python 和 REST 使用 Azure OpenAI On Your Datahttps://learn.microsoft.com/en-us/azure/foundry-classic/openai/references/on-your-data
将 Azure OpenAI 与 Pinecone 向量数据集成https://learn.microsoft.com/en-us/azure/foundry-classic/openai/references/pinecone
使用 Azure OpenAI 语音将文本转换为语音https://learn.microsoft.com/en-us/azure/foundry-classic/openai/text-to-speech-quickstart
使用 Azure OpenAI Whisper 转录语音https://learn.microsoft.com/en-us/azure/foundry-classic/openai/whisper-quickstart
使用 Microsoft Foundry SDK 构建 RAG 聊天应用https://learn.microsoft.com/en-us/azure/foundry-classic/tutorials/copilot-sdk-build-rag
部署
主题URL
评估 Foundry Models 的部署选项https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/deployments-overview
规划 Microsoft Foundry 的组织级部署https://learn.microsoft.com/en-us/azure/foundry-classic/concepts/planning
选择 Foundry 模型区域和部署类型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/concepts/models-sold-directly-by-azure-region-availability
使用 Azure CLI 和 Bicep 部署 Foundry 模型https://learn.microsoft.com/en-us/azure/foundry-classic/foundry-models/how-to/create-model-deployments
使用 Bicep 部署 Microsoft Foundry 中心https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-azure-ai-hub-template
使用 Terraform 配置 Foundry 中心和项目https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/create-hub-terraform
在 Foundry 中使用托管计算部署模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/deploy-models-managed
将 Hugging Face Hub 模型部署到 Foundry 托管终结点https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/deploy-models-managed-hugging-face
将 Foundry Models 部署到托管的按需付费计算资源https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/deploy-models-managed-pay-go
在 Foundry 中将模型部署为无服务器 APIhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/deploy-models-serverless
在 GitHub Actions 管道中运行 Foundry 评估https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/evaluation-github-action
在 Foundry 中使用托管计算部署微调模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/fine-tune-managed-compute
通过 Foundry 无服务器 API 部署微调模型https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/fine-tune-serverless
将 Foundry 提示流部署为托管联机终结点https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/flow-deploy
在 Azure 上部署和操作 Agent Framework 工作流https://learn.microsoft.com/en-us/azure/foundry-classic/how-to/how-to-deploy-migrated-agent-framework-workflow
将 Azure OpenAI 资源升级到 Foundryhttps://learn.microsoft.com/en-us/azure/foundry-classic/how-to/upgrade-azure-openai
按区域检查 Microsoft Foundry 功能支持情况https://learn.microsoft.com/en-us/azure/foundry-classic/reference/region-support

评论

支持 Markdown