How does the cost optimization work in this skill?

It routes simple tasks like data inspection to the low-cost Claude Haiku model, while reserving complex reasoning and coding for Sonnet or Opus, resulting in significant cost savings compared to single-model approaches.

What kind of files can I analyze with this skill?

DS-STAR supports a wide variety of data formats including CSV, JSON, and plain text, and can process both structured datasets and unstructured information.

What is DS-STAR for Claude Code?

DS-STAR is a multi-agent framework that uses a 'Structured Thought and Action' architecture to automate data science tasks like exploratory analysis and code execution directly within your development environment.

Does DS-STAR handle code errors automatically?

Yes, the framework includes a dedicated Debugger agent that identifies execution errors and automatically generates fixes through iterative refinement rounds.

Are the analysis results reproducible?

Absolutely. Every run creates a structured directory containing the exact prompts used, the generated Python scripts, and comprehensive execution logs for full transparency.

DS-STAR Data Science Framework

Name: DS-STAR Data Science Framework
Author: Token-Eater

byToken-Eater

0•

Data Science & ML

Automates complex data science workflows using a multi-agent architecture and optimized model routing for efficient, iterative data analysis.

DS-STAR (Data Science - Structured Thought and Action) is a sophisticated multi-agent framework designed to transform raw data into actionable insights within Claude Code. By orchestrating seven specialized agents—ranging from data analyzers to Python coders and verifiers—it manages the entire data science lifecycle including planning, execution, and iterative debugging. The framework optimizes performance and cost by dynamically routing tasks to the most suitable Claude model (Haiku, Sonnet, or Opus), providing a reproducible and highly efficient alternative to manual exploratory data analysis.

Key Features

01Comprehensive artifact generation that saves prompts, code, and logs for full research reproducibility

02Cost-optimized model routing across Haiku, Sonnet, and Opus to reduce analysis expenses by up to 60%

03Iterative refinement cycles with automated debugging and verification gates for high-accuracy results

04Multi-agent orchestration using specialized roles for data analysis, strategy planning, and coding

05Native support for diverse data formats including CSV, JSON, and unstructured text files

060 GitHub stars

Use Cases

01Automating exploratory data analysis (EDA) and generating statistical summaries for new datasets

02Performing cost-effective data processing workflows using mixed-model logic for large datasets

03Executing iterative time-series or correlation analysis with automated Python code generation and debugging

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add token-eater/skills-marketplace ds-star

For use in Claude.ai and ChatGPT

Download Skill