Guide · Agent skill
How to use the fabric-lakehouse agent skill for Microsoft Fabric
Updated · written with AI from the project's own documentation and checked against our catalogue data
The fabric-lakehouse skill is a set of instructions that gives an AI agent background knowledge about Microsoft Fabric Lakehouse, including its folders, Delta tables, schemas, shortcuts, access control and performance tuning. This guide explains what an agent skill is, how to install this one, how to ask for help in plain language, and what to expect. By the end you will know how to set the skill up and when it is useful.
Fabric Lakehouse Guidegithub/awesome-copilot · MIT Download the skill Quick start
- 1Download the ZIP. Download the fabric-lakehouse ZIP file from this page.
- 2Unzip the skill. Unzip the file so the fabric-lakehouse folder with SKILL.md is available on your computer.
- 3Place the folder. Move the folder to ~/.claude/skills/ for personal use, or to .claude/skills/ inside a project.
- 4Or upload in the Claude apps. In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- 5Ask a Fabric Lakehouse question. Describe your Lakehouse question in plain language and mention Fabric so the agent uses the right context.
What an agent skill is and what this one does
An agent skill is a folder containing a SKILL.md file and optional reference files. When installed, an AI agent can read these files to learn context for a specific topic before it answers you or helps with a task.
The fabric-lakehouse skill is a knowledge guide. According to its description, it offers descriptions of Lakehouse data components, organization with schemas and shortcuts, access control, and code examples, to support designing, building and optimizing Lakehouse solutions. It teaches the agent context; it does not perform actions in your Fabric environment.
- Explains the Tables and Files folders and how Delta tables, CSV and Parquet are stored.
- Describes schemas (the default schema is dbo), schema shortcuts and other shortcut types.
- Covers workspace roles, OneLake security, and column-level and row-level security.
- Describes table optimization with OPTIMIZE (with optional Z-ordering) and cleanup with VACUUM.
- Points to PySpark examples for reading and writing data and to guidance on getting data into a Lakehouse.
How to install the fabric-lakehouse skill
Download the ZIP file for the skill from this page and unzip it. Then place the skill folder in one of two locations, depending on how you want to use it.
For a personal setup across projects, put it in your user skills folder at ~/.claude/skills/. For a single project, put it in that project's .claude/skills/ folder. If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal use: ~/.claude/skills/
- Single project: the .claude/skills/ folder inside the project
- Claude apps: Settings → Capabilities → Skills, then upload the skill
How to trigger the skill with natural requests
You do not need a special command. Ask the agent a question about Fabric Lakehouse in plain language, and it can use the skill when the request matches the skill's description. The skill's own 'When to Use' list covers requests to explain Lakehouse concepts, design or optimize a solution, learn the components, or manage tabular and non-tabular data.
Be specific about your situation. Mention Fabric and Lakehouse by name so the agent does not confuse it with other platforms.
- Explain how the Tables and Files folders differ in a Fabric Lakehouse.
- Review this Lakehouse design and suggest where schemas would help.
- Which role should a analyst have in a Fabric workspace to read Lakehouse data?
- Show me a PySpark example for writing a Delta table in a Lakehouse.
- How often should I run OPTIMIZE and VACUUM on my Lakehouse tables?
What to expect and what it needs
The skill lists no required tools or packages. It does not run scripts and does not use the network, so it works as plain instructions and reference text. The skill also contains a PySpark reference and a data-ingestion reference, which the agent reads when your question needs them.
The guidance describes Microsoft Fabric as it is documented in this skill. Fabric features change over time, so treat the answers as a starting point and check Microsoft documentation before making production decisions.
- Requires: nothing listed
- Runs scripts: no
- Uses network: no
- Files: SKILL.md, references/pyspark.md, references/getdata.md
Example PySpark starting point from the skill
The PySpark reference includes a Spark configuration example that enables Fabric optimizations, along with reading and writing patterns. The snippet below is taken from that reference. Check it against your own environment before running it, since it writes to a Delta table.
spark.conf.set("spark.sql.parquet.vorder.enabled", "true")
spark.conf.set("spark.microsoft.delta.optimizeWrite.enabled", "true")Tips for getting useful answers
Ask for one decision at a time, such as a schema layout or a security approach, and give the agent the context it needs, like how many workspaces you have or whether data comes from Azure, AWS or Google Cloud. The skill covers ADLS Gen2, Amazon S3, Dataverse and Google Cloud Storage shortcuts, so naming the source helps.
Use the agent's answers to shape your design, then confirm details in Microsoft's Fabric documentation.
- Name the source system when asking about shortcuts.
- Ask for the reasoning behind a recommendation, not just the answer.
- Check any command or setting in official documentation before running it.
What people use it for
- Explaining Tables versus Files and when to use Delta tables in a Fabric Lakehouse to a new team member.
- Reviewing a Lakehouse design for schemas, shortcuts and access control before a build starts.
- Getting a starting PySpark or Spark SQL example for reading or writing Lakehouse tables.
- Planning a table maintenance routine that uses OPTIMIZE and VACUUM.
Frequently asked questions
Is the fabric-lakehouse skill free to use?
The skill is released under the MIT licence, as listed on this page. Check the licence file in the source repository for the full terms.
What do I need to install the fabric-lakehouse skill?
The skill lists no required tools. You download the ZIP, unzip it, and place it in ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
Does the fabric-lakehouse skill connect to my Fabric workspace?
No. It is a knowledge guide and does not run scripts or use the network. It gives the agent context, but it does not change anything in your Fabric environment.
Is the skill safe to run?
The skill runs no scripts and uses no network, according to its listing. The PySpark examples write data when you run them, so review and test any code in a non-production workspace first.
Can I use this skill for plain Databricks or Snowflake?
No. The skill is for Microsoft Fabric Lakehouse, and its advice will not apply to non-Fabric platforms such as plain Databricks or Snowflake.
How is this different from reading the Microsoft documentation myself?
The skill gives an agent a structured summary to answer your questions in context. Because Fabric changes, the official Microsoft documentation remains the authority for current behaviour.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you are a data engineer designing or reviewing a Microsoft Fabric Lakehouse, an analyst new to Fabric, or a developer who wants PySpark or Spark SQL starting points. Pick something else if you need guaranteed-current Fabric behaviour or you work on a non-Fabric platform, since the content is a summary to verify against Microsoft documentation.
Download the skill