agentleFS
alifadl009alifadl009/microsoft-fabric-claude-skill/AGENTS.md

microsoft-fabric-claude-skill

AGENTS.md3 starsChanged 4 months agoMITView on GitHub

Update Check: At session start, check for skills-for-fabric updates by reading the remote package.json version from https://github.com/microsoft/skills-for-fabric (via git fetch origin main --quiet && git show origin/main:package.json or GitHub API with authentication) and comparing with the local package.json version. Show changelog if update available. You are an AI assistant specialized in Microsoft Fabric development. All Fabric operations require Azure AD authentication. For development:

The file

We split this AGENTS.md into the rules it gives an agent

8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All Fabric operations require Azure AD authentication.

    line 15
  2. Distinctive:

    Use Medallion architecture: Bronze (raw) → Silver (cleaned) → Gold (aggregated)

    line 52
  3. Distinctive:

    Use Delta Lake for Lakehouse tables

    line 96
  4. Very specific:

    Use has over contains for indexed string search in KQL

    line 98
  5. Very specific:

    Use .create-merge table and .create-or-alter function for idempotent KQL schema deployment

    line 99
  6. Distinctive:

    Use alphanumeric PascalCase names (3–63 chars) for Eventstream nodes

    line 101
  7. Distinctive:

    Use SQL operator for CDC Debezium payload flattening in Eventstreams

    line 102
  8. Distinctive:

    Use Activator skills for Reflex item definitions, rule templates, and action payloads

    line 103

≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Microsoft Fabric Development Agent

Update Check: At session start, check for skills-for-fabric updates by reading the remote package.json version from https://github.com/microsoft/skills-for-fabric (via git fetch origin main --quiet && git show origin/main:package.json or GitHub API with authentication) and comparing with the local package.json version. Show changelog if update available.

You are an AI assistant specialized in Microsoft Fabric development.

Architecture Mode

  • This repository uses a hybrid model: Agents → Skills → Common.
  • For cross-workload orchestration (medallion architecture, migration, ETL across Spark + SQL + KQL), use agents/FabricDataEngineer.agent.md.
  • Delegate endpoint-specific implementation depth to skills in skills/.

Authentication

All Fabric operations require Azure AD authentication. For development:

# Login to Azure
az login

# Get token for Fabric REST API
az account get-access-token --resource https://api.fabric.microsoft.com

# Get token for SQL connections (Warehouse, Lakehouse SQL Endpoint)
az account get-access-token --resource https://database.windows.net

Primary Reference

Fabric REST APIs: https://learn.microsoft.com/en-us/rest/api/fabric/articles/

Workload Documentation

Key Patterns

Data Architecture

  • Use Medallion architecture: Bronze (raw) → Silver (cleaned) → Gold (aggregated)
  • Lakehouse for data engineering, Warehouse for SQL analytics
  • Delta Lake format for all Lakehouse tables
  • For Materialized Lake View SQL authoring and incremental refresh optimization, use skills/spark-authoring-cli/SKILL.md and its MLV resource documents.
  • For MLV refresh scheduling, job monitoring, and cancellation, use skills/mlv-operations-cli/SKILL.md.

Development

  • PySpark with mssparkutils for notebooks
  • T-SQL with surface area limitations for Warehouse
  • SQL Database in Fabric (OLTP): T-SQL DDL/DML, read-only queries, and performance diagnostics (see sqldb-authoring-cli, sqldb-consumption-cli, sqldb-operations-cli skills)
  • KQL for real-time analytics (always use time filters)
  • Power Query M for Dataflows Gen2 transformations (see dataflows-authoring-cli and dataflows-consumption-cli skills)
  • Eventstream for real-time event ingestion (graph-based topology with sources, operators, destinations)
  • Activator for Reflex alerts, notifications, and automated actions over Fabric events and data
  • DAX for Semantic Model measures
  • Semantic model development (see semantic-model-authoring)
  • Power BI report planning skill: skills/powerbi-report-planning/SKILL.md — requirements, page plan, approval gate
  • Power BI report design skill: skills/powerbi-report-design/SKILL.md — archetype routing, layout, theme, accessibility
  • Power BI report authoring skill: skills/powerbi-report-authoring/SKILL.md — PBIR/PBIP file mechanics, Desktop reload/screenshot
  • Power BI report management skill: skills/powerbi-report-management/SKILL.md — Fabric report item CRUD via az rest
  • Spark operations skill: skills/spark-operations-cli/SKILL.md — read-only triage for failed jobs, stuck sessions, performance bottlenecks

Operations

  • REST APIs for programmatic management
  • Pipelines for orchestration
  • Parameterize everything for reusability
  • Warehouse operations skill: skills/sqldw-operations-cli/SKILL.md — performance diagnostics, slow queries, query insights

Activator / Reflex

  • Authoring skill: skills/activator-authoring-cli/SKILL.md — create Activator items, sources, rules, conditions, and actions
  • Consumption skill: skills/activator-consumption-cli/SKILL.md — inspect Activator definitions, rules, sources, and actions

Power BI / FabricIQ

  • Consumption skill: skills/semantic-model-consumption/SKILL.md — raw DAX queries against semantic models via MCP ExecuteQuery tool
  • FabricIQ skill: skills/fabriciq/SKILL.md — multi-step Power BI data analysis (discover, inspect, resolve, generate, execute)
  • ⚠️ MANDATORY: Before calling any FabricIQ MCP tool, read skills/fabriciq/SKILL.md in full (see agents/FabricIQ.agent.md § Pre-Flight).

Fabric IQ / Ontology (preview)

  • Authoring skill: skills/fabriciq-ontology-authoring-cli/SKILL.md — define entity types, properties (incl. timeseries), relationship types, and bind them to lakehouse/Eventhouse tables via the item-definition REST API
  • Consumption skill: skills/fabriciq-ontology-consumption-cli/SKILL.md — read ontology items for agent grounding context and route ontology-backed queries to the matching per-datasource consumption skill

Constraints

Must

  • Use Delta Lake for Lakehouse tables
  • Include time filters in KQL queries (where Timestamp > ago(...))
  • Use has over contains for indexed string search in KQL
  • Use .create-merge table and .create-or-alter function for idempotent KQL schema deployment
  • Discover KQL Database query URI via Fabric REST API before connecting
  • Use alphanumeric PascalCase names (3–63 chars) for Eventstream nodes
  • Use SQL operator for CDC Debezium payload flattening in Eventstreams
  • Use Activator skills for Reflex item definitions, rule templates, and action payloads
  • Handle secrets via Key Vault or environment variables
  • Validate T-SQL features against supported surface area

Avoid

  • Hardcoded IDs or connection strings
  • SELECT * on large tables without LIMIT
  • Unbounded streaming queries
  • Complex calculated columns in Semantic Models (use measures)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.