Generic prompts are a poor substitute for a repeatable Fabric workflow. A skill gives an AI coding assistant focused instructions, the relevant tools, safe ordering, and a definition of done.
This deck introduces Skills for Fabric and shows where they fit across data engineering, analytics, administration, application development, and migration.
Watch the interactive deck
The deck is hosted on Roesli, so it does not require a GitHub or Deckio account.
What is a skill?
A skill is a reusable package of instructions for a focused task. It can define:
- when the workflow applies;
- which Fabric tools and endpoints matter;
- the safe sequence of operations;
- what the result should contain; and
- which checks must pass before the work is complete.
That matters in Fabric because a single outcome can cross Lakehouses, Warehouses, notebooks, pipelines, semantic models, Power BI, OneLake security, deployment, and capacity management.
Start with one repeatable task
Choose a workflow with a clear input and output. Good candidates include:
- reviewing a semantic model and its measures;
- creating a medallion Lakehouse structure;
- checking pipeline or notebook deployment readiness;
- assessing a Synapse or Databricks migration; or
- reviewing Fabric capacity, governance, or workspace configuration.
Write down the inputs, expected output, tools, safety constraints, and verification steps. That short contract is the useful part of the skill.
The aim is not just faster code generation. It is consistent work with visible assumptions and explicit validation. When the workflow improves, update the skill instead of relying on somebody’s memory of the last successful run.
Loading comments…