microsoft/skills-for-fabric
16 skills751 installs6.8k starsGitHub
Install
npx skills add https://github.com/microsoft/skills-for-fabricSkills in this repo
1Powerbi Authoring Clipowerbi-authoring-cli is a reference skill for Microsoft Fabric and Power BI semantic models. Solo builders and small teams use it when an agent must retrieve model metadata, full TMDL definitions, or refresh status without guessing which API owns each field. The guide explains why properties are split across Fabric Items and Power BI Datasets endpoints, and points to refresh and schedule operations documented in the main Fabric skills package. It fits anyone automating dataset audits, CI checks on semantic models, or agentic workflows that download definitions and inspect refresh health. You still need proper API authentication and workspace access configured per the parent skill. It does not replace the full authoring CLI workflow—it accelerates correct API choice and reduces failed calls during build-time analytics integration work.71installs2Spark Authoring CliSpark Authoring CLI is a Microsoft Fabric skill resource that encodes PySpark data-engineering patterns for solo builders shipping lakehouse workloads. It emphasizes explicit schemas for production ingestion, Delta Lake as the default table format, quality checks at boundaries, lineage metadata columns, error handling with logging, MERGE-based incremental updates, and sensible partitioning. Guidance also steers you away from inferSchema in prod, toward batch unless streaming is required, and toward ingesting real files from lakehouse paths when URLs are supplied. Use it while building Fabric notebooks or jobs and when hardening pipelines before production operations. It pairs naturally with Fabric deployment and monitoring practices elsewhere in the journey.68installs3E2e Medallion Architecturee2e-medallion-architecture teaches agents how to stand up a full Medallion lakehouse on Microsoft Fabric: separate lakehouse tiers, ingestion into Bronze, cleansing and conformation in Silver, and analytics-ready Gold tables. It targets solo data builders and indie SaaS teams who need a repeatable pattern instead of ad-hoc notebooks. Expect PySpark and Delta Lake specifics, pipeline wiring between layers, and Spark configuration notes tuned per tier. The skill insists on a one-time-per-session update check (check-updates or package version compare) and documents Fabric workspace/item lookup patterns so agents do not guess IDs. Use when triggers mention bronze silver gold, multi-layer lakehouse, or e2e data pipeline on Fabric—not for generic SQL CRUD apps outside the Fabric ecosystem.64installs4Powerbi Consumption CliPower BI semantic model discovery (powerbi-consumption-cli) is a Microsoft skills-for-fabric agent skill that documents read-only DAX patterns for exploring semantic model metadata through INFO.VIEW.* and INFO.* rowsets. Solo and indie builders who ship on Microsoft Fabric or Power BI—whether a SaaS dashboard, internal ops reporting, or a client analytics deliverable—can paste these queries into a consumption CLI or query endpoint to inventory tables, columns, measures, and relationships without mutating the model. The readme prioritizes a small scope-estimation probe (four object counts) so an agent can judge how heavy a full metadata pass will be before running wider INFO.VIEW scans. High-value columns are spelled out per rowset: calculated tables, storage mode, hidden flags, lineage tags, measure state, KPI linkage, and relationship cross-filter semantics. Use it during model design, handoff documentation, or pre-refactor audits when you need a consistent, repeatable metadata snapshot the agent can reason over in chat.63installs5Spark Consumption Clispark-consumption-cli is a Microsoft skills-for-fabric agent workflow for interactive lakehouse analytics through Fabric’s Lakehouse Livy API and PySpark or Spark SQL. Solo builders and small data teams reach for it when plain SQL is not enough—DataFrame pipelines, joins across lakehouses, Delta time-travel, and semi-structured JSON need a Python Spark session instead of a single query tool. The skill encodes Fabric topology, workspace and item discovery with JMESPath filtering, and CLI consumption patterns from the shared COMMON-CORE references. First use in a session must run the check-updates path so package guidance stays current. Triggers are intentionally narrow: user language mentioning PySpark, Livy, Spark DataFrames, or Python-based lakehouse analysis. If someone only wants lightweight SQL, the skill itself says to pick a different path. Expect intermediate Fabric literacy, live cloud workspaces, and credentials managed outside the skill text.59installs6Sqldw Authoring CliSQLDW Authoring CLI Quick Reference equips solo builders and agents operating Microsoft Fabric warehouses to execute T-SQL authoring from the terminal using sqlcmd. It standardizes connection variables against *.datawarehouse.fabric.microsoft.com endpoints, then documents high-frequency patterns: table DDL, CTAS with explicit types, ALTER and DROP, bulk INSERT…SELECT, file-based upserts, and TRUNCATE-style maintenance. The skill points to shared CORE authoring guidance and script templates for longer workflows, so you do not duplicate policy across repos. Use it when you are provisioning fact tables, staging loads, or validating row counts before a dashboard ships. It assumes Azure AD authenticated sqlcmd (-G) rather than legacy SQL auth snippets. Intermediate familiarity with T-SQL and Fabric workspace naming helps; beginners can still copy patterns once a warehouse endpoint exists.58installs7Check Updatescheck-updates is a meta skill for Microsoft’s skills-for-fabric marketplace. Solo builders who depend on Fabric-tuned agent capabilities use it to avoid silently running stale skill packs: on first invocation of any marketplace skill in a given week, it reads shared session state, compares the installed version to GitHub releases, and surfaces what changed when an upgrade is available. State lives in a user-level JSON file under ~/.config/fabric-collection so the check is shared across sessions and plugins, not repeated every chat turn. If the last check for this plugin is fewer than seven UTC calendar days ago, the skill exits quickly so you stay productive. It fits anytime in the journey you touch Fabric skills—before research in Idea, before deployment helpers in Ship, or during Operate when you audit tooling—but the maintenance story is fundamentally about keeping procedural knowledge current as the marketplace evolves.54installs8Sqldw Consumption Clisqldw-consumption-cli is a Microsoft skills-for-fabric agent skill that compresses how solo builders and small data teams interact with Fabric SQL data warehouse through sqlcmd from bash or PowerShell. It standardizes connection variables (FABRIC_SERVER, FABRIC_DB), Azure Active Directory authentication via -G, and export-oriented flags so query results land in clean CSV or TSV files without manual cleanup. The skill pairs CLI ergonomics with consumption-oriented monitoring patterns documented in the broader Fabric SQLDW consumption core guide, and it steers agents toward script-templates for repeatable jobs like filtered fact extracts. Expect prerequisite checks for sqlcmd installation (including winget install hints) and az login before any run. It is aimed at operators and builders who already have a warehouse endpoint and need fast, copy-paste-safe automation rather than GUI-only exploration—ideal when you are scheduling extracts, validating row counts after ETL, or watching warehouse load between releases.53installs9Eventhouse Authoring CliEventhouse Authoring CLI is a template skill for solo builders and small teams standing up Kusto tables inside Microsoft Fabric Eventhouse. It packages copy-paste Bash and PowerShell flows that wrap Azure CLI az rest calls to the cluster management API, so you can define schemas, wire CSV ingestion mappings, and kick off blob ingest without memorizing KQL management syntax. The scripts use a small run_mgmt helper that serializes db and csl into JSON, authenticates with the Kusto resource, and surfaces results via jq. You supply cluster URI, database name, and optionally a blob path for ingest. It fits builders who already use Fabric for observability or product analytics and want repeatable, reviewable infra scripts instead of one-off portal clicks. It is not a full IaC replacement or a query authoring tutor—it accelerates the first table-and-ingest pass on Eventhouse.52installs10Eventhouse Consumption Clieventhouse-consumption-cli packages KQL schema discovery patterns for Microsoft Fabric Eventhouse consumption scenarios. Solo builders and small teams use it when an agent must explore what exists in a cluster before writing analytics queries—listing tables, pulling CSL or JSON schema, estimating cardinality with column statistics, and enumerating functions or materialized views by folder. The skill reads like a field manual of .show and profiling idioms rather than a deploy pipeline, which suits iterative analytics and data QA on live Eventhouse data. It pairs with Fabric skills stacks when you need trustworthy table and column context before building dashboards or downstream automations. Expect KQL fluency and an attached Eventhouse workspace; it does not replace governance approval for production schema changes.50installs11Best Practices Checkbest-practices-check is a meta-quality skill for maintainers of Microsoft Fabric agent skills. When you add or refresh skills such as spark-consumption-cli, sqldw-authoring-cli, or medallion architecture packs, this skill tells the agent to pull current official and community Fabric guidance, line it up against what your SKILL.md actually says, and call out missing patterns or stale recommendations. It is not a substitute for running workloads in Fabric; it is editorial assurance that procedural knowledge in the repo still matches what Microsoft documents today. Solo builders who ship custom Fabric skills for their own agents use it to avoid shipping misleading CLI steps or incomplete medallion guidance. The normalization table resolves ambiguous names like spark or SQL endpoint to the right skill file. Treat it as a release gate companion to human review, especially when Fabric features move quickly.31installs12Quality CheckQuality-check is a checker skill for Microsoft’s skills-for-fabric repo that runs a local Python workflow before you commit skill changes. Solo builders and small teams contributing Fabric-oriented agent skills use it after adding or editing folders under skills/ to catch structural violations, semantic conflicts between similar skills, broken cross-references, and content issues that would fail upstream automation. You need Python 3.8+ with PyYAML and requests, then invoke the checker from the repository root on Windows or Unix. The run surfaces critical versus warning counts and persists detail in quality-report.json so you can fix blockers before opening a pull request. It is narrowly scoped to this repository’s layout and checker script—not a generic linter for arbitrary SKILL.md trees elsewhere.28installs13Skill Testskill-test is the meta-skill for Microsoft’s skills-for-fabric evaluation framework. Contributors and maintainers use it when they need structured eval plans, visibility into what tests exist, fresh datasets, pass/fail metrics, or a gap analysis of skills without coverage. The skill parses natural-language intent—phrases like add tests, show eval results, or test coverage—and routes to the correct workflow instead of guessing file locations. Execution is explicitly anchored in the repo’s tests/ directory so agents do not invent one-off scripts. It fits solo builders extending Fabric agent skills who treat evals as part of shipping reliable procedural knowledge, not an afterthought. Pair it with individual domain skills once those skills need regression harnesses.26installs14Sqldw Operations Clisqldw-operations-cli is a query reference skill for Microsoft Fabric Data Warehouse operations. It gives agents copy-ready T-SQL against the built-in queryinsights schema and system DMVs so solo builders and small data teams can answer “what is slow?” and “what is expensive?” without reinventing monitoring SQL. Queries cover long-running workload summaries, top resource consumers, and related performance analyses documented in the skill body. Historical context is bounded: queryinsights retains roughly thirty days of history and results may lag completion by up to fifteen minutes—facts agents should cite when interpreting dashboards. Use it in Operate when a Fabric SQL warehouse is already deployed and you need repeatable diagnostic phrasing for Claude Code or similar agents, not for greenfield schema design.26installs15Search Consumption Clisearch-consumption-cli is a Microsoft Fabric agent skill for solo builders and small teams who manage many workspaces and need to find notebooks, lakehouses, semantic models, or other items without guessing workspace IDs first. It drives the Fabric Catalog Search and related CLI patterns so your agent can answer "where is X?" or "list all items of type Y" and emit stable IDs for the next API step. The skill is explicit that search targets items, not workspace names, and it cross-links workspace resolution via the common CLI docs. It expects you to run the pack’s update check once per session so command shapes stay aligned with the published skills repo. Use it during integration scripting, incident triage when an item name is known but location is not, and whenever you are chaining discovery into deployment or refresh automation rather than browsing the Fabric portal manually.24installs16Synapse MigrationSynapse-migration is a Microsoft Skills for Fabric reference that helps solo data builders estimate Fabric capacity SKUs when moving from Azure Synapse Spark pools. The material is explicitly a planning aid: the migration workflow itself assumes the target workspace already has assigned capacity, but you still need sane SKU guidance so Spark jobs do not starve Warehouse or Power BI on the same tenant. Tables map common Synapse pool shapes—Small through XXL node counts—to Fabric F-skus with vCore totals, typical dataset sizes, and concurrent job headroom. That supports indie analytics products and internal BI stacks where one person owns both the Spark notebooks and the monthly Azure bill. Invoke it when scoping a Synapse-to-Fabric cutover, rightsizing after a lift-and-shift, or explaining to stakeholders why F32 versus F64 matters for medium ETL. Pair it with Fabric deployment skills for execution steps beyond sizing alone.24installs