Why does it prefer Mart models over Staging models?

Marts contain cleaned, transformed, and business-ready logic, whereas Staging models are closer to raw data and may lack the necessary transformations for accurate business reporting.

How does it handle complex questions the Semantic Layer can't answer?

It uses Approach 2: fetching the compiled SQL for a metric and then applying custom logic (like CASE statements or extra joins) before executing it against the warehouse.

Can I use this skill to run 'dbt build' or 'dbt test'?

No, this skill is strictly for answering data questions. It is not intended for development workflows like building, testing, or validating dbt models.

Does this skill require a dbt Semantic Layer to work?

No, while it prioritizes the dbt Semantic Layer for accuracy, it can fall back to model discovery and analyzing project manifest files if a semantic layer is not configured.

dbt Natural Language Data Querying

Name: dbt Natural Language Data Querying
Author: dbt-labs

bydbt-labs

•

分析与监控

Answers complex business questions by intelligently navigating dbt semantic layers, models, and project artifacts.

关于

This skill transforms Claude into a specialized data analyst for dbt projects. It utilizes a sophisticated four-tier hierarchy to resolve natural language questions: first attempting to query the dbt Semantic Layer, then modifying compiled SQL for custom needs, discovering relevant Mart models, and finally analyzing local manifest artifacts. This approach ensures high-accuracy data retrieval while maintaining best practices by prioritizing defined business metrics over raw staging tables. It is ideal for teams who want to bridge the gap between their dbt transformation layer and immediate business insights.

主要功能

Proactive suggestions for improving Semantic Layer definitions based on query gaps
Hierarchical query resolution (Semantic Layer > SQL Modification > Model Discovery > Manifest Analysis)
Intelligent Mart model prioritization to ensure business logic consistency
Artifact-based analysis using manifest.json and catalog.json for local projects
Automated SQL refinement based on compiled dbt metric logic
13 GitHub stars

使用场景

Identifying missing dimensions or metrics in a dbt project's semantic configuration
Answering aggregate business questions like 'What were total sales last quarter by region?'
Investigating specific customer segments without writing manual SQL joins

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dbt-labs/dbt-agent-skills answering-natural-language-questions-with-dbt

For use in Claude.ai and ChatGPT

Download Skill

GitHub

关于