Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Azure Agent Skills is a first-party, Microsoft-authored catalog of AI-readable skills for Azure development. Each skill is a structured index of authoritative Microsoft Learn content, packaged so AI coding assistants can discover and use it to provide accurate, up-to-date guidance on Azure services.
The skills follow the Agent Skills open standard and work with any compatible AI coding assistant, including GitHub Copilot, Claude Code, Cursor, Gemini CLI, Codex CLI, and others.
Use cases
- Give AI coding assistants expert-level Azure guidance grounded in official documentation
- Improve the accuracy and completeness of agent responses for Azure development tasks
- Enable developers to get best practices, troubleshooting flows, architecture patterns, and security guidance directly in their editor
- Provide structured, up-to-date knowledge that complements live data tools like MCP servers
How Agent Skills work
Agent Skills are lightweight, filesystem-based capability modules. Each skill is a folder containing a SKILL.md file with YAML frontmatter (name, description, compatibility) and structured content organized by category (Troubleshooting, Best Practices, Architecture, Security, Configuration, and more).
Skills use progressive disclosure to load content efficiently:
- Discovery — The AI assistant reads the skill's
nameanddescriptionfrom YAML frontmatter to determine relevance - Instructions — When triggered, the assistant loads the full
SKILL.mdcontent, which contains curated documentation links organized by topic - Fetch — The assistant retrieves the latest content from Microsoft Learn using the Learn MCP Server or a web fetch tool
This architecture separates the knowledge index (what to look up and when) from the content delivery (fetching the latest docs at query time), ensuring skills stay current as documentation evolves.
The skill catalog is automatically generated and maintained by a purpose-built pipeline that scans Azure product documentation, identifies high-value procedural and decision-making content, and transforms it into structured skill files. The pipeline continuously evaluates for upstream content changes and produces incremental updates.
Skill catalog
The repository contains 193 skills covering Azure's core services across these categories:
| Category | Skills | Examples |
|---|---|---|
| Infrastructure & Management | 27 | Azure Resource Manager, Backup, Monitor, Policy, Cost Management, Arc |
| Specialized | 27 | Quantum, Maps, VMware Solution, SAP, Health Data Services, Virtual Desktop |
| Networking | 24 | Virtual Network, Front Door, Firewall, VPN Gateway, ExpressRoute, Load Balancer |
| AI & ML | 20 | AI Services, Machine Learning, Document Intelligence, Speech, Bot Service |
| Data & Analytics | 20 | Cosmos DB, SQL Database, Data Factory, Synapse Analytics, Databricks |
| Security & Identity | 17 | Key Vault, RBAC, Sentinel, Defender for Cloud, Confidential Computing |
| DevOps | 14 | Azure DevOps, Pipelines, Container Registry, Deployment Environments, Dev Box |
| Compute | 13 | Functions, Container Apps, Virtual Machines, Kubernetes Service, App Service |
| Integration | 12 | Logic Apps, API Management, Event Grid, Service Bus, Event Hubs |
| Storage | 9 | Blob Storage, Files, Elastic SAN, NetApp Files, Container Storage |
| IoT & Edge | 7 | IoT Hub, IoT Central, IoT Edge, IoT Operations, Digital Twins |
Role-based curated bundles are available for common scenarios (Essentials, Infrastructure, Data Engineering, AI/ML, Security, DevOps).
The complete skill catalog is available at docs/CATALOG.md.
Requirements
- Network access is required. Skills reference Microsoft Learn documentation URLs that the AI assistant fetches at query time.
- A compatible AI coding assistant that supports the Agent Skills open standard (
SKILL.mdformat). - Recommended: The Microsoft Learn MCP Server for optimal documentation retrieval. If the Learn MCP Server is not available, skills fall back to web fetch capabilities built into most AI assistants.
No authentication is required to use the skills themselves. The skills are static files that run locally in your development environment.
Getting started
Installation
Clone the repository:
git clone https://github.com/MicrosoftDocs/agent-skills.gitCopy the skill folders from
skills/to your AI assistant's skills directory:AI Assistant Project-level path Global path GitHub Copilot {project}/.github/skills/~/.copilot/skills/Claude Code {project}/.claude/skills/~/.claude/skills/Codex CLI {project}/.codex/skills/~/.codex/skills/Cursor {project}/.cursor/skills/— Gemini CLI {project}/.gemini/skills/— Important
Copy the folders inside
skills/(e.g.,azure-functions/,azure-storage/), not theskills/folder itself.VS Code users: Enable Chat: Use Agent Skills (
chat.agent.skills) in Settings.Start asking your AI assistant Azure questions. Skills activate automatically based on your query.
Usage
Once installed, ask your AI assistant questions naturally:
- "Help me set up Azure Functions with Durable Functions"
- "What are the best practices for Azure Container Apps?"
- "I'm getting throttling errors on my storage account — how do I diagnose this?"
The assistant will automatically discover relevant skills, load the structured guidance, and fetch the latest documentation from Microsoft Learn to provide accurate, grounded answers.
Compatibility
These skills work with any AI coding assistant that supports the Agent Skills specification:
| AI Coding Assistant | Type | Skills Path |
|---|---|---|
| GitHub Copilot | VS Code Extension | .github/skills/ |
| Claude Code | Anthropic CLI | .claude/skills/ |
| Codex CLI | OpenAI CLI | .codex/skills/ |
| Cursor | AI-native IDE | .cursor/skills/ |
| Gemini CLI | Google DeepMind | .gemini/skills/ |
| Antigravity IDE | DeepMind IDE | .agent/skills/ |
| OpenCode | Open-source CLI | .agent/skills/ |
Relationship to the Learn MCP Server
Agent Skills and the Learn MCP Server are complementary:
- Learn MCP Server provides tools — search, fetch, and code sample retrieval against Microsoft Learn content
- Agent Skills provide knowledge — structured guidance on what to look up, when, and how to apply it
Skills use the Learn MCP Server (or equivalent fetch capability) as their content delivery mechanism. Together, they give AI assistants both the judgment to identify what's relevant and the ability to retrieve the latest authoritative content.
Availability and pricing
Azure Agent Skills is publicly available as an open-source repository under dual license:
- Documentation content: Creative Commons Attribution 4.0 International
- Code: MIT License
There is no charge to use the skills.
Limitations
- Skills require network access to fetch documentation from Microsoft Learn. Without network access, skills provide curated link indexes but cannot retrieve full content.
- The skill catalog currently covers 193 Azure services. Coverage is expanding with each pipeline run as new documentation is published.
- Skills are generated from publicly available Microsoft Learn documentation. Content from training modules, learning paths, and instructor-led courses is not included at this time.
- The generation pipeline runs incrementally. There may be a short delay between documentation updates on Microsoft Learn and corresponding skill updates.
How to provide feedback
- GitHub Issues: https://github.com/MicrosoftDocs/agent-skills/issues (preferred for bugs, gaps, or feature requests)
- Contributions: The repository welcomes pull requests. See the contribution guidelines in the README.