mcpField
M
microsoft-docs

microsoft-docs

github.com/microsoftDevelopment3.2k120 installsUpdated today
Repository

About microsoft-docs

A documentation research skill for finding Microsoft Learn material and retrieving complete pages. It supports focused queries across concepts, tutorials, configuration, limits, and best practices, with an `mslearn` CLI alternative.

Key features

  1. 01Search Microsoft documentation
  2. 02Fetch full documentation pages
  3. 03Guidance for focused queries
  4. 04CLI alternative with `mslearn`

Use cases

  1. 01Looking up service concepts and architecture
  2. 02Finding tutorials and configuration details
  3. 03Checking service limits or best practices

MCP support

It uses the `microsoft_docs_search` and `microsoft_docs_fetch` tools to search and retrieve Microsoft documentation through MCP. If the Learn MCP server is unavailable, it documents equivalent `mslearn` CLI commands.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/microsoft/skills --skill microsoft-docs

After installing, type/microsoft-docsin Claude Code to confirm it loaded.

Details

From the project README

Microsoft Docs

Tools

ToolUse For
microsoft_docs_searchFind documentation—concepts, guides, tutorials, configuration
microsoft_docs_fetchGet full page content (when search excerpts aren't enough)

When to Use

  • Understanding concepts — "How does Cosmos DB partitioning work?"
  • Learning a service — "Azure Functions overview", "Container Apps architecture"
  • Finding tutorials — "quickstart", "getting started", "step-by-step"
  • Configuration options — "App Service configuration settings"
  • Limits & quotas — "Azure OpenAI rate limits", "Service Bus quotas"
  • Best practices — "Azure security best practices"

Query Effectiveness

Good queries are specific:

# ❌ Too broad
"Azure Functions"

# ✅ Specific
"Azure Functions Python v2 programming model"
"Cosmos DB partition key design best practices"
"Container Apps scaling rules KEDA"

Include context:

  • Version when relevant (.NET 8, EF Core 8)
  • Task intent (quickstart, tutorial, overview, limits)
  • Platform for multi-platform docs (Linux, Windows)

When to Fetch Full Page

Fetch after search when:

  • Tutorials — need complete step-by-step instructions
  • Configuration guides — need all options listed
  • Deep dives — user wants comprehensive coverage
  • Search excerpt is cut off — full context needed

Why Use This

  • Accuracy — live docs, not training data that may be outdated
  • Completeness — tutorials have all steps, not fragments
  • Authority — official Microsoft documentation

CLI Alternative

If the Learn MCP server is not available, use the mslearn CLI from the command line instead:

# Run directly (no install needed)
npx @microsoft/learn-cli search "azure functions timeout"

# Or install globally, then run
npm install -g @microsoft/learn-cli
mslearn search "azure functions timeout"
MCP ToolCLI Command
microsoft_docs_search(query: "...")mslearn search "..."
microsoft_docs_fetch(url: "...")mslearn fetch "..."

The fetch command also supports --section <heading> to extract a single section and --max-chars <number> to truncate output.

Comments

Markdown supported