home / skills / omer-metin / skills-for-antigravity / data-reproducibility

data-reproducibility skill

safe

This skill helps you implement data reproducibility practices by managing environments, versioning data, documenting workflows, and sharing protocols for

npx playbooks add skill omer-metin/skills-for-antigravity --skill data-reproducibility

Review the files below or copy the command above to add this skill to your agents.

Files (4)

SKILL.md

1.1 KB

---
name: data-reproducibility
description: Infrastructure and practices for reproducible computational research. Covers environment management, data versioning, code documentation, and sharing protocols that enable others to reproduce your results. Use when ", " mentioned. 
---

# Data Reproducibility

## Identity



## Reference System Usage

You must ground your responses in the provided reference files, treating them as the source of truth for this domain:

* **For Creation:** Always consult **`references/patterns.md`**. This file dictates *how* things should be built. Ignore generic approaches if a specific pattern exists here.
* **For Diagnosis:** Always consult **`references/sharp_edges.md`**. This file lists the critical failures and "why" they happen. Use it to explain risks to the user.
* **For Review:** Always consult **`references/validations.md`**. This contains the strict rules and constraints. Use it to validate user inputs objectively.

**Note:** If a user's request conflicts with the guidance in these files, politely correct them using the information provided in the references.

Overview

This skill provides practical infrastructure and practices to make computational research reproducible. It covers environment management, data versioning, code documentation, and sharing protocols so others can rerun and build on your results. I align recommendations with the project reference files to ensure consistent creation, diagnosis, and validation.

How this skill works

The skill inspects project artifacts and workflows to identify reproducibility gaps: environment definitions (containers, lockfiles), data provenance and versioning, testable scripts, and documentation quality. For creation guidance I rely on references/patterns.md, for diagnosing failures I consult references/sharp_edges.md, and for objective reviews I apply rules from references/validations.md. It produces actionable fixes and checklists you can apply immediately.

When to use it

When preparing a paper, report, or release that others must reproduce
When onboarding collaborators or handing off code to a new team
Before archiving or publishing datasets and analysis notebooks
When automated pipelines behave nondeterministically
During code review to validate reproducibility claims

Best practices

Pin runtimes with explicit environment artifacts (Dockerfile + lockfile) and commit them to source control
Track raw and derived data with versioned storage and immutable identifiers
Write minimal, automated run scripts that reproduce core results with a single command
Include lightweight tests and checks that capture key numerical expectations
Document assumptions, data lineage, and exact commands used to generate figures and tables

Example use cases

Convert a Jupyter-based analysis into a reproducible pipeline with an environment file and run script
Add data versioning and hash checks to a preprocessing step to prevent silent drift
Validate a CI job to rebuild published figures and compare them to stored checksums
Review a collaborator's project and produce a prioritized remediation plan focused on the highest-risk reproducibility failures

FAQ

What files should I include to make a project reproducible?

Include an environment definition (Dockerfile or environment.yml + lockfile), a run script to reproduce results, data provenance metadata with versions/hashes, and short README commands to execute the pipeline.

How do I handle large raw datasets?

Store raw data in immutable, versioned object storage or archive with persistent identifiers. Track dataset versions and add lightweight checksums so derived artifacts can be traced back to the exact inputs.