What is the benefit of using the tdx agent skill?

It allows you to manage AI agents as code, enabling version control, local editing of prompts, and structured configuration of tools and knowledge bases.

How do I test my agents locally?

You can use the 'tdx chat' command followed by the agent path to initiate a terminal-based conversation and verify agent behavior before pushing to production.

How do knowledge bases work in Treasure Data agents?

Knowledge bases can be table-based (querying TD databases) or text-based (Markdown files), allowing agents to retrieve specific information during a conversation.

Which Claude models are supported?

The skill supports various Claude models, including Claude 3.5 Sonnet and Haiku, with specific settings for reasoning effort and temperature.

What is the @ref syntax used for?

The @ref syntax is a standardized way to create cross-resource references between agents, knowledge bases, and prompts within your YAML configuration.

Treasure Data Agent Development

Name: Treasure Data Agent Development
Author: treasure-data

bytreasure-data

•

데이터 과학 및 ML

Develops and manages sophisticated LLM agents within the Treasure Data ecosystem using a YAML-based configuration workflow.

소개

The tdx agent skill streamlines the creation and management of AI agents on the Treasure Data platform by enabling a local-first development workflow. It provides a structured approach to defining agent behavior, integrating knowledge bases, and configuring tools like web search or image generation through simple YAML and Markdown files. By supporting commands for pulling, pushing, and testing agents directly from the command line, it allows developers to maintain version control over their AI projects while leveraging advanced features like extended reasoning and cross-resource referencing.

주요 기능

6 GitHub stars
Seamless integration with table-based and text-based knowledge bases.
Local-first workflow for pulling and pushing agent configurations to Treasure Data.
Built-in CLI chat interface for rapid testing and iteration of agent responses.
Declarative YAML configuration for agent behavior, models, and temperature settings.
Advanced tool support including web search, image generation, and multi-agent delegation.

사용 사례

Creating multi-agent systems where a coordinator agent delegates tasks to specialized SQL experts.
Deploying production-ready AI agents with specific reasoning effort levels for complex problem solving.
Building customer support agents that query Treasure Data databases for personalized context.

소개

주요 기능

6 GitHub stars
Seamless integration with table-based and text-based knowledge bases.
Local-first workflow for pulling and pushing agent configurations to Treasure Data.
Built-in CLI chat interface for rapid testing and iteration of agent responses.
Declarative YAML configuration for agent behavior, models, and temperature settings.
Advanced tool support including web search, image generation, and multi-agent delegation.

사용 사례

Creating multi-agent systems where a coordinator agent delegates tasks to specialized SQL experts.
Deploying production-ready AI agents with specific reasoning effort levels for complex problem solving.
Building customer support agents that query Treasure Data databases for personalized context.