td-skills

🥈Silver

Treasure Data Skills for Claude Code enables AI-assisted development on the Treasure Data platform, providing curated skills for Digdag workflows, scheduling, operators, and runtime behavior.

1740Updated 1mo ago

Intermediate30min to implementautomation

Saves ~60 min per use

Quick InstallView Source

git clone https://github.com/treasure-data/td-skills.git

Works with:

Claude

Overview

About This Skill

td-skills is an official Treasure Data skill pack for Claude Code that equips AI coding agents with accurate, platform-specific knowledge for building on Treasure Data. The skill set covers Digdag workflow authoring, operator usage (including TD-specific caveats for s3_move>, loop>, and py>), scheduling options, SLA configuration, retry semantics, secret expansion, and system concurrency limits. Skills are organized into focused reference files—operators, scheduling, and runtime—so the agent can answer questions and generate correct code without requiring users to manually consult documentation. The project follows a stable/next release channel model, making it suitable for teams that need predictable, versioned AI assistance in production Treasure Data environments.

How to Use

Install the skill pack via the .claude-plugin directory into your Claude Code environment. Skills are versioned across stable and next channels; maintainers can use scripts/release.sh to tag prereleases or promote to stable. Once installed, Claude Code will reference the bundled operator, scheduling, and runtime documentation when assisting with Treasure Data projects.

Use Cases

Authoring Digdag workflows with correct idempotent ETL patterns (split-task DELETE+INSERT with group _retry)

Configuring workflow schedules, SLA settings, and timezone options on Treasure Data

Understanding operator authentication, output variables, and Secrets Reference for TD platform operators

Getting accurate retry semantics, variable behavior, and concurrency limits when building TD pipelines

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/treasure-data/td-skills

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Act as a data engineer using Treasure Data Skills for Claude Code. I need to [TASK] using [DATA] from [DATA_SOURCE]. Provide the exact TD Skills code to accomplish this, including any necessary setup or configuration.

Example Output

# Treasure Data Skills for Claude Code Solution

## Task: Customer Segmentation
## Data Source: Salesforce CRM

```sql
-- Setup: Ensure TD Skills is connected to your Salesforce account
CONFIGURE CONNECTION salesforce_connection
  WITH
  endpoint = 'https://yourinstance.salesforce.com'
  , username = 'your_username'
  , password = 'your_password'
  , security_token = 'your_token'
;

-- Extract customer data from Salesforce
CREATE OR REPLACE TABLE salesforce_customers AS
SELECT
  id
  , name
  , email
  , total_purchases
  , last_purchase_date
FROM salesforce.sobject.Account
WHERE is_active = true
;

-- Segment customers based on purchase behavior
CREATE OR REPLACE TABLE customer_segments AS
SELECT
  *
  , CASE
    WHEN total_purchases > 1000 THEN 'High Value'
    WHEN total_purchases BETWEEN 500 AND 1000 THEN 'Medium Value'
    ELSE 'Low Value'
  END AS segment
FROM salesforce_customers
;

-- Export segments to a CSV file for marketing team
EXPORT TABLE customer_segments
TO '/output/customer_segments.csv'
WITH HEADER = true
;
```