ds-star

Coordinate seven AI agents to analyze data, generate code, and verify results.

5|Updated Nov 15, 2025
One-click install
npx skills add https://github.com/Token-Eater/skills-marketplace --skill ds-star
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ds-star
Source: https://github.com/Token-Eater/skills-marketplace/tree/main/skills/ds-star
Command: npx skills add https://github.com/Token-Eater/skills-marketplace --skill ds-star

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

DS-STAR automates and orchestrates multi-agent data science workflows to handle data analysis, coding, verification, and presentation within Claude Code, reducing manual overhead and enabling reproducible research.

Core Features & Use Cases

  • Multi-agent orchestration with dedicated roles: Analyzer, Planner, Coder, Verifier, Router, Debugger, Finalyzer for end-to-end data science tasks.
  • Multi-model optimization and cost savings through routing tasks to Haiku, Sonnet, and Opus.
  • Reproducible pipelines with artifact saving and resume support, capable of iterative refinement for complex queries.

Quick Start

Invoke the ds-star skill to analyze your data with seven specialized agents and return a structured final result.

Frequently Asked Questions about ds-star

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

FAQPage Schema
How do I automate multi-agent data science workflows in Claude Code?

Multi-agent data science workflows in Claude Code are automated by coordinating seven specialized AI agents to handle data analysis, code generation, verification, and final presentation. DS-STAR routes tasks to cost-efficient models and saves artifacts for reproducible research.

Can I optimize data analysis costs by routing tasks to different Claude models?

Data analysis costs are optimized by routing specific tasks to different Claude models like Haiku, Sonnet, and Opus. DS-STAR assigns code generation and verification steps to the most cost-efficient model available while maintaining robust results.

How do I ensure reproducibility for exploratory data analysis and data cleaning pipelines?

Reproducibility for exploratory data analysis and data cleaning pipelines is ensured by saving artifacts and supporting resume functionality. DS-STAR stores intermediate results to allow iterative refinement and verification across multiple data formats.

Do I need Python and pandas to run automated data analysis agents?

Python and pandas are required to run automated data analysis agents in the Claude Code environment. DS-STAR relies on these dependencies to execute generated code, verify results, and handle structured data formats.

What is the best way to validate generated code results in automated data analysis?

Generated code results in automated data analysis are validated through a dedicated Verifier and Debugger agent. DS-STAR includes optional debug steps to check executable code outputs and ensure robust, structured final answers.

Why does multi-agent data analysis fail to produce reproducible results without artifact management?

Multi-agent data analysis fails to produce reproducible results without artifact management because intermediate outputs are lost. DS-STAR prevents this by saving artifacts during data cleaning and code generation to support resume and iterative refinement.