Does this skill perform the actual data processing?

No, it focuses on building the specification document. It defines the 'what' and 'how' of the analysis so that the implementation can be executed consistently.

What is the Analysis Spec Builder skill?

It is a Claude Code skill that helps researchers create formal specification documents for physics data analysis, ensuring all technical details and tools are correctly defined.

Can I use custom datasets with this skill?

Yes. While it defaults to PHYSLITE for ATLAS physics, it is designed to incorporate any specific dataset identifiers or analysis names you provide.

What tools does this skill recommend by default?

By default, it suggests using ServiceX and func_adl for data fetching, and awkward, hist, and vector for data manipulation, alongside pyhf for statistical analysis.

Physics Analysis Spec Builder

Name: Physics Analysis Spec Builder
Author: gordonwatts

bygordonwatts

0•

Data Science & ML

Generates and refines structured physics analysis specifications using standardized templates and domain-specific best practices.

The Analysis Spec Builder is a specialized skill for Claude Code designed to help researchers and data scientists formalize their physics analysis workflows. It transforms high-level requests—such as requests for specific plots or dataset explorations—into detailed, technical specification documents. By utilizing a standardized template, the skill ensures that critical details like dataset identifiers, selection criteria, histogram definitions, and statistical tools (such as ServiceX, func_adl, and cabinetry) are explicitly defined. It follows an iterative process, asking focused follow-up questions to fill gaps, ultimately producing a production-ready specification.md file for collaborative research.

Key Features

01Support for high-energy physics standards like PHYSLITE datasets

02Template-driven specification generation for scientific research

03Iterative refinement through targeted follow-up questions

04Integration of standardized tool recommendations (ServiceX, func_adl, pyhf)

05Automatic placeholder replacement and markdown file export

060 GitHub stars

Use Cases

01Drafting complex data selection and fitting requirements for PHYSLITE datasets

02Formalizing a quick physics analysis idea into a technical specification document

03Standardizing workflow steps and histogram definitions for a research team

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add gordonwatts/analysis-specification analysis-spec-builder

For use in Claude.ai and ChatGPT

Download Skill