fabric-api-discovery - SKILL.md Agent Skill

name: fabric-api-discovery description: > Discover Fabric APIs, OpenAPI specs, item schemas, and best practices using the Fabric MCP Server. Use when exploring available Fabric workloads, looking up API specifications, finding item definition formats, or managing OneLake files programmatically. All MCP tools run locally for reference.

Fabric API Discovery

Fabric MCP Server

The Fabric Pro-Dev MCP Server provides local-first API discovery and OneLake management without connecting to live Fabric environments.

Setup

npm install -g fabric-pro-dev-mcp-server

Requires Node.js 18+.

MCP Configuration

Add to your .mcp.json or agent MCP config:

{
  "mcpServers": {
    "fabric-pro-dev": {
      "command": "npx",
      "args": ["-y", "fabric-pro-dev-mcp-server"]
    }
  }
}

Power BI MCP (Live Queries)

For live DAX queries against semantic models:

{
  "mcpServers": {
    "PowerBIQuery": {
      "type": "http",
      "url": "https://api.fabric.microsoft.com/v1/mcp/powerbi",
      "oauthClientId": "<MICROSOFT_PUBLIC_CLIENT_ID_FROM_CURRENT_DOCS>",
      "oauthPublicClient": true
    }
  }
}

Use the Microsoft-documented public client ID from the current Fabric MCP documentation rather than hardcoding a GUID in this repository.

API Discovery Tools

Tool	Purpose	When to Use
`publicapis_list`	List all Fabric workload types with APIs	Starting Fabric development, exploring workloads
`publicapis_get`	Get OpenAPI spec for a specific workload	Building integrations, checking request/response schemas
`publicapis_platform_get`	Get platform-level API specs	Working with core/admin APIs
`publicapis_bestpractices_get`	Get best practices documentation	Before implementing API patterns
`publicapis_bestpractices_examples_get`	Get request/response examples	Need concrete API call samples
`publicapis_bestpractices_itemdefinition_get`	Get item schema definitions	Creating or updating item definitions

OneLake Tools

Tool	Purpose
`onelake download file`	Download files from OneLake
`onelake upload file`	Upload files to OneLake
`onelake file list`	List files in OneLake path
`onelake file delete`	Delete files from OneLake
`onelake directory create`	Create OneLake directories
`onelake directory delete`	Delete OneLake directories
`onelake item list`	List workspace items
`onelake item list-data`	List items via DFS endpoint
`onelake item create`	Create new Fabric items

Available Workloads

Workloads with public API specifications include:

Lakehouses
Warehouses
Notebooks
Spark Job Definitions
Data Pipelines
Semantic Models (Power BI)
KQL Databases
Eventhouse
Real-Time Intelligence (Eventstreams)
ML Models & Experiments
Reports
Environments

Item Definitions

Fabric items can be created/updated via definition-based workflows using base64-encoded file parts:

POST /v1/workspaces/<wsId>/items/<itemId>/updateDefinition[?updateMetadata=true]
{
  "definition": {
    "format": "<format>",
    "parts": [
      { "path": "<filePath>", "payload": "<base64>", "payloadType": "InlineBase64" }
    ]
  }
}

Use publicapis_bestpractices_itemdefinition_get to discover the expected format and parts for each item type.

Gotcha: Omitting ?updateMetadata=true silently ignores the .platform part.

Developer vs Consumer Patterns

Developers (REST + Protocol)

Use REST APIs to create/manage artifacts
Use protocol-specific connections for data access:
- Spark/PySpark for Lakehouse data
- ODBC/JDBC for Warehouse queries
- XMLA/DAX for Semantic Models
- KQL for Real-Time Intelligence

Consumers (MCP)

Use MCP servers for natural-language queries
Supported for: Semantic Models, Warehouses, Lakehouse SQL Endpoints
No ODBC/JDBC setup needed — MCP handles connections

CLI Tooling

Tool	Install	Purpose
Fabric CLI	`pip install ms-fabric-cli`	Command-line Fabric operations
fabric-cicd	`pip install fabric-cicd`	CI/CD automation for Fabric items
Azure CLI	`az rest --resource https://api.fabric.microsoft.com`	REST API calls with auto-auth