Does this skill overwrite my existing Treasure Data descriptions automatically?

No. The skill follows a strict human-in-the-loop workflow. It generates descriptions into a CSV file for your review and requires explicit confirmation after a dry-run before any data is written back to Treasure Data.

Do I need a separate Anthropic API key to use the generation feature?

The skill supports two modes: API mode using an ANTHROPIC_API_KEY for automated generation, and a Claude Code mode that provides manual instructions if you prefer to use the chat interface.

What happens if a write-back error occurs during the process?

The skill includes a built-in rollback mechanism. It takes 'before' snapshots of your table schemas, allowing you to restore Treasure Data to its original state if a failure occurs.

Can I use this for PII (Personally Identifiable Information) detection?

Yes. During the schema extraction and generation phases, the AI identifies potential PII based on column names and sample data, flagging them in the review CSV for your verification.

Treasure Data Dictionary Helper

Name: Treasure Data Dictionary Helper
Author: treasure-data

bytreasure-data

•

Gestión de Bases de Datos

Automates the generation and maintenance of Treasure Data column descriptions using AI-powered schema analysis and human-in-the-loop review.

This skill streamlines the tedious process of maintaining data dictionaries within Treasure Data by orchestrating a complete workflow from schema extraction to metadata write-back. It uses Claude to automatically generate 80-90% of column descriptions based on business context and sample data, providing a structured CSV-based review process for human validation before publishing changes back to source tables. It is an essential tool for data engineers and analysts looking to improve data discoverability, governance, and AI-readiness within their Customer Data Platform.

Características Principales

01Automated AI-powered column description generation with 80-90% coverage

02Structured CSV export for human review and PII validation

03Context-aware generation tailored to specific industries and use cases

04End-to-end orchestration from schema extraction to TD write-back

0516 GitHub stars

06Safety-first write-back with dry-run previews and automated rollbacks

Casos de Uso

01Identifying and flagging PII across complex CDP table structures

02Cleaning up legacy data dictionaries to improve self-service analytics

03Rapidly documenting new Treasure Data segments for downstream users

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add treasure-data/td-skills data-dictionary-helper

For use in Claude.ai and ChatGPT

Download Skill