Give your AI coding assistant Microsoft Fabric guidance for building data workflows, querying data, and migrating workloads.
Skills for Fabric gives AI coding assistants reusable instructions for Microsoft Fabric APIs, queries, and operational workflows. Describe your task in natural language, and the assistant selects matching skills to guide its work across data engineering, analytics, administration, and migration. The skills provide instructions; the assistant carries out the actions through available tools and APIs.
Microsoft distributes the collection as two separate plugins: fabric-skills for Fabric workloads and powerbi-authoring for Power BI authoring. In GitHub Copilot CLI, you can also name an experimental agent specialization in your prompt to focus on administration, data engineering, application development, or Power BI questions.
When raw data needs to become reporting-ready tables, ask for a Bronze, Silver, and Gold architecture. The architecture skill guides ingestion, PySpark and Delta transformations, pipeline orchestration, and the handoff to a semantic model.
Use the warehouse skill to investigate failed or canceled queries, slow execution, and capacity usage spikes. It guides read-only diagnostics with Query Insights and Capacity Metrics correlation.
Convert Databricks notebooks and jobs, Synapse Spark or SQL workloads, HDInsight Spark and Hive code, or Synapse pipeline definitions with the corresponding migration skill. Use the capacity estimation skill when you need a sizing and pricing estimate before the move.
Build an Eventstream to route incoming events, use Eventhouse for KQL analysis, and define Activator rules when a threshold should trigger a Teams or email action.
Ask FabricAdmin to document a named workspace. Its workflow inspects notebooks, pipelines, SQL logic, and semantic models, then writes an overview and reports by artifact type as Markdown.
With the separate powerbi-authoring plugin, plan a report from an existing semantic model, design its pages, edit PBIR definitions in a PBIP project, validate the result, and publish when requested. Local report authoring requires a PBIP project using the PBIR format.
The Fabric bundle includes the agents, Fabric workload skills, and hosted MCP connections. Power BI report authoring and the local modeling MCP server come from the separate powerbi-authoring plugin; semantic-model-authoring appears in both plugin manifests.
Showing all 25
Plan and build Bronze, Silver, and Gold data layers with orchestration.
Author and run Spark notebooks, investigate failures, and manage Materialized Lake Views.
Create and inspect Dataflows Gen2, edit Power Query M, and upgrade Gen1 dataflows.
Manage remote Project Osmos data engineering tasks; direct portal Copilot interaction is unsupported.
Author and query warehouses, inspect SQL endpoints, and investigate warehouse performance.
Query and change Fabric SQL databases; investigate blocking and query regressions.
Build and inspect Eventstream sources, transformations, destinations, and routing.
Create, update, inspect, and delete event type and payload schema definitions.
Create Eventhouse database objects, configure ingestion, and query data with KQL.
Create and inspect threshold rules that trigger Teams or email actions.
Bring Azure Monitor telemetry into Eventhouse and correlate it with business data.
Answer business questions using existing Power BI reports and semantic models.
Model ontology entities, relationships, and data bindings; explore schemas and graphs.
Build and manage semantic models, DAX measures, relationships, refresh, and deployment.
Plan, design, edit, validate, and publish reports; requires the separate Power BI plugin.
Estimate capacity costs and compare SKUs and billing options before migration.
Convert Databricks notebooks, jobs, storage paths, and utilities for Fabric.
Convert Synapse Spark, Lake Database, and Dedicated SQL Pool workloads for Fabric.
Convert HDInsight Spark, Hive, and Oozie workloads to Fabric equivalents.
Convert Synapse pipelines, linked services, datasets, and parameters for Fabric Data Factory.
Connect workspaces to Git, commit or pull changes, and resolve synchronization conflicts.
Manage shared variables and value sets for workspaces and deployment stages.
Configure deployment stages and promote selected items between development, test, and production.
Find Fabric items across workspaces; Dataflow Gen1 and Gen2 are unsupported.
Audit catalog governance, apply guarded remediation, and curate item metadata and table or column descriptions.
No matching results.
Coordinate workspace administration, governance, capacity planning, observability, and documentation.
Install the Fabric bundle and restart GitHub Copilot CLI to load it. Name FabricAdmin in your prompt and identify the workspace, for example: “Using FabricAdmin, document my workspace Finance.”
Its agent definition describes the administration and workspace documentation workflows.
Coordinate data engineering across Spark, SQL, pipelines, migration, and lakehouse architecture.
Install the Fabric bundle and restart GitHub Copilot CLI to load it. Name FabricDataEngineer in your prompt and describe the data sources, target workspace, and desired result, such as a lakehouse with Bronze, Silver, and Gold layers.
Its agent definition explains how it coordinates workload-specific skills.
Build applications that access Fabric data through ODBC, XMLA, and REST APIs.
Install the Fabric bundle and restart GitHub Copilot CLI to load it. Name FabricAppDev in your prompt, identify the Fabric data your application should use, and describe the application you want to build.
Its agent definition covers application connectivity and local development.
Discover Power BI artifacts and answer business questions by querying their data.
Install the Fabric bundle and restart GitHub Copilot CLI to load it. Name FabricIQ in your prompt and ask a business question about a Power BI report or semantic model you can access. Include the report link or name when you have it.
Its agent definition describes artifact discovery and the query workflow.
Discover Power BI reports and semantic models, inspect their metadata, and query their data.
Inspect and edit Power BI semantic models through the hosted authoring service.
Run T-SQL against Fabric warehouses and SQL analytics endpoints.
Inspect and edit Power BI semantic models through a locally launched server.
Add Microsoft's Skills for Fabric marketplace, then install fabric-skills and powerbi-authoring for complete coverage. Each plugin installs as a complete bundle.
Install both plugins from inside a Claude Code session. The Power BI plugin's local modeling server runs through Node.js, so have Node.js available. For live Fabric tasks, you need access to the target workspace, permission for the requested operations, and an Azure CLI sign-in. Publishing and editing Power BI content requires a Power BI tenant and the corresponding permissions.
Add the Skills for Fabric marketplace in your Claude Code session.
/plugin marketplace add microsoft/skills-for-fabricInstall the complete fabric-skills plugin. If Claude Code opens the plugin details panel, choose the installation scope and confirm.
/plugin install fabric-skills@fabric-collectionInstall the complete powerbi-authoring plugin for Power BI reports and semantic models, confirming its installation in the plugin panel.
/plugin install powerbi-authoring@fabric-collectionQuit and restart Claude Code to load the installed skills.
/skillsIn Claude Code's installed-skills list, check that the Fabric skills and the Power BI report skill are available.
Use an installed GitHub Copilot CLI with Node.js and PowerShell 7. For live Fabric tasks, you also need access to the target workspace, permission for the requested operations, and an Azure CLI sign-in. Publishing and editing Power BI content requires a Power BI tenant and the corresponding permissions.
Add the Skills for Fabric marketplace inside your GitHub Copilot CLI session.
/plugin marketplace add microsoft/skills-for-fabricInstall the complete fabric-skills plugin.
/plugin install fabric-skills@fabric-collectionInstall the complete powerbi-authoring plugin for Power BI reports and semantic models.
/plugin install powerbi-authoring@fabric-collectionQuit and restart GitHub Copilot CLI to load the installed skills.
/skillsIn GitHub Copilot CLI's installed-skills list, check that the Fabric skills and the Power BI report skill are available.