genie-metadata-optimizer

Analyze evaluation failures and generate metadata patch proposals for Genie Space.

5|6|Updated Jan 26, 2026
One-click install
npx skills add https://github.com/databricks-solutions/vibe-coding-workshop-template --skill genie-metadata-optimizer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: genie-metadata-optimizer
Source: https://github.com/databricks-solutions/vibe-coding-workshop-template/tree/main/data_product_accelerator/skills/semantic-layer/genie-optimization-workers/03-genie-metadata-optimizer
Command: npx skills add https://github.com/databricks-solutions/vibe-coding-workshop-template --skill genie-metadata-optimizer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses situations where evaluation scores for AI-generated data insights are below target, indicating a need to refine the underlying metadata that guides the AI.

Core Features & Use Cases

  • Metadata Optimization: Generates proposals to improve table and column descriptions, metric view definitions, and TVF parameters.
  • Failure Analysis: Analyzes evaluation failures to identify root causes and suggest specific metadata changes.
  • Use Case: After running benchmarks, if Genie consistently fails to identify the correct table for a specific business question, this Skill can analyze the failures and propose adding a business_definition or synonyms to the relevant table's metadata to guide Genie more effectively.

Quick Start

Use the genie-metadata-optimizer skill to generate metadata change proposals based on recent evaluation failures.

Frequently Asked Questions about genie-metadata-optimizer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I optimize Genie Space metadata when AI insights evaluation scores are below target?

To optimize Genie Space metadata, analyze recent evaluation failures to identify root causes, then generate targeted patch proposals to refine UC table descriptions, column descriptions, and metric views for better AI insights.

Why does Genie fail to identify the correct UC table for specific business questions?

Genie may fail to identify the correct UC table when metadata lacks sufficient business context. Adding a business definition or synonyms to the relevant table metadata guides the AI more effectively toward the correct data source.

What is the best way to fix evaluation failures in Databricks Genie Spaces?

The best way to fix Genie evaluation failures is applying ASI-grounded introspection or GEPA to analyze the failures and generate specific metadata change proposals for affected UC tables, metric views, and TVF parameters.

Can I use ASI-grounded introspection to improve TVF parameters in a Genie Space?

Yes, you can use ASI-grounded introspection (L1) or GEPA (L2) to analyze Genie evaluation failures and generate targeted patch proposals that specifically improve TVF parameters and metric view definitions within your semantic layer.

What metadata elements can be refined to improve Databricks Genie AI insights?

You can refine UC table descriptions, column descriptions, metric view definitions, and TVF parameters. Adding business definitions and synonyms to metadata helps guide Genie more effectively when generating AI data insights.