Developers increasingly use AI coding agents to describe intent, generate code, and iterate solutions. However, general-purpose assistants do not automatically understand Fabric-specific APIs, authentication patterns, item discovery, SQL endpoints, or deployment practices.
Skills for Fabric closes that gap. The repository provides reusable, open-source instructions that help GitHub Copilot, Claude Code, Cursor, Windsurf, and other compatible AI coding tools to work more accurately with Fabric workloads.
Skills for Fabric now supports SQL database in Fabric, adding workload-specific guidance for authoring, consuming, and operating SQL databases as part of broader Fabric solutions. These skills help AI coding agents understand Fabric-specific APIs, authentication patterns, item discovery, SQL endpoints, and operational workflows so they can support complete solutions.
Why skills matter
Skills give AI coding agents access to specialized, workload-specific guidance only when it is relevant to the task. Rather than embedding lengthy instructions in every conversation, an agent begins with a concise description of each available skill and loads the detailed guidance only when a request matches it. This keeps prompts focused, reduces unnecessary context usage, and enables agents to support a broader set of capabilities without sacrificing relevance.
Skills for Fabric now supports SQL database in Fabric, allowing agents to reason across complete Fabric solutions rather than focus on isolated database tasks. Modern Fabric solutions often span data pipelines, lakehouses, warehouses, notebooks, semantic models, and operational SQL databases. The SQL database skill connects these components in a reviewable implementation plan, reducing the setup work required to identify the appropriate APIs, authentication requirements, and operational patterns.
For example, a developer can ask an agent to create an operational SQL database for a property management application, define tables and relationships, and prepare the data for downstream analytics. The agent can identify the required Fabric items, propose the implementation sequence, and show how the operational data connects to reporting or AI experiences. This gives the developer one reviewable plan instead of separate instructions for each workload.
What SQL database skills support
SQL database support extends Skills for Fabric across three common stages of solution development: authoring operational data structures, consuming data through queries and applications, and operating database resources safely within Fabric.
Author
- Create SQL databases in Fabric workspaces as part of broader solution designs.
- Generate tables, keys, relationships, indexes, triggers, and sample data with reviewable implementation steps.
- Build AI-ready schemas with vector columns for semantic search, retrieval, and application scenarios.
Consume
- Explore database schemas and discover available objects.
- Generate and run SQL queries to access and work with data.
- Use SQL database results in applications, analytics workflows, and connected Fabric experiences.
Operate
- Discover database items, identifiers, SQL endpoints, and connection details.
- Manage database resources through Fabric APIs and CLI commands.
- Validate schema and data changes while applying review and approval of safeguards.
The SQL database skill produces reviewable implementation steps, surfaces assumptions, and identifies validation points before execution. This planning-first approach lets users inspect and approve proposed changes before they affect connected Fabric workloads or shared resources.
Depending on the request, the agent can return a solution outline, proposed schema definitions, SQL queries, API or CLI commands, and validation steps tailored to the Fabric environment. The skill connects operational data design with downstream analytics and application requirements rather than treating database operations in isolation. Users can then review the proposed implementation sequence, confirm assumptions, and validate dependencies before execution.
This approach helps teams understand how SQL database resources fit within a larger Fabric solution and makes it easier to evaluate changes that may affect connected workloads. By presenting a reviewable plan before execution, the skill supports a workflow that combines AI-assisted productivity with human oversight, helping developers move from intent to implementation with greater confidence.
Getting started
To get started, choose a compatible AI coding tool, install Skills for Fabric, and sign in to your Fabric environment with your existing identity and permissions. Then describe the SQL database task you want to complete in natural language. The agent uses the installed skill to apply the relevant APIs, commands, query patterns, and operational guidance.
GitHub Copilot CLI and Visual Studio Code
Install from the GitHub plugin marketplace:
/plugin install fabric-skills@copilot-plugins
Or install a focused bundle:
/plugin install fabric-authoring@copilot-plugins
/plugin install fabric-consumption@copilot-plugins
/plugin install fabric-operations@copilot-plugins
GitHub Copilot CLI plugin installation is the recommended path. For other AI coding tools, the repository ships root-level configuration files that load automatically when the repo is opened in your project.
Claude Code, Cursor, and Windsurf
Clone the Skills for Fabric repository into your project or add it as a submodule alongside your existing code. Claude Code, Cursor, and Windsurf then load the included Fabric guidance from the repository configuration files.
git clone https://github.com/microsoft/skills-for-fabric.git
- Claude Code loads CLAUDE.md and the included .mcp.json configuration.
- Cursor loads .cursorrules.
- Windsurf loads .windsurfrules.
Authentication
Most Fabric operations require Microsoft Entra authentication. Sign in and request an access token for the Fabric API:
az login
az account get-access-token --resource https://app.fabric.microsoft.com
The installed skills provide workload-specific guidance for requesting and using the appropriate token audiences.
Example prompt:
Help me design a Fabric solution for property management analytics. Use a SQL database in Fabric for operational entities such as properties, tenants, leases, and payments. Recommend how the database should connect with downstream analytics workloads, generate the initial schema, identify required relationships and indexes, and create a validation plan before implementation.
For more details on how to use SQL database skills, see Work with SQL database skills in Microsoft Fabric.
What’s next
The Skills for Fabric public library is continuing to grow. With SQL database now included, users can start applying Fabric skills to more complete solution patterns that span operational data, analytics, and AI-ready application scenarios.
Install the appropriate Skills for Fabric bundle, open your Fabric project, and try the example prompt with your own database scenario. To request support for another workload, scenario, or migration path, open an issue in the repository. Community skill contributions are welcome.
Explore the Skills for Fabric catalog to find guidance for other Fabric workloads.
Microsoft Employee