maxwell2732

maxwell2732 / paper-replicate-agent-demo

Public

A structured Claude Code workflow for empirically replicating published research using UK Biobank (UKB) data, created by Chen Zhu 朱晨 | 遗传社科研究.

47
35
100% credibility
Found Feb 23, 2026 at 19 stars 2x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
HTML
AI Summary

A structured workflow using AI to empirically replicate findings from published epidemiology papers with UK Biobank data, producing validation reports on result matches.

How It Works

1
📚 Discover the replication kit

You hear about a handy starter kit that helps everyday researchers check if published studies' results hold up with real data.

2
📁 Set up your workspace

Download the kit and create simple folders to organize your research papers and datasets.

3
📄 Add your paper and data

Drop the research paper PDF and your approved dataset into the right folders.

4
🤖 Ask the AI helper to replicate

Chat with your built-in AI assistant and say 'replicate this paper' – it reads everything and gets to work like a smart research partner.

5
🔍 Watch the magic unfold

The assistant plans the steps, runs the analyses, compares results to the paper's claims, and notes any differences.

6
📊 Get your validation report

Receive a clear report detailing what matched perfectly, what was close, and clear explanations for any gaps.

🎉 Replication complete!

You now have a trustworthy summary to build on, cite, or share, making your own research stronger and more reliable.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 47 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is paper-replicate-agent-demo?

This repo sets up a Claude Code workflow to replicate empirical research papers using UK Biobank data. Drop in a paper PDF and your data; an agent powered by Anthropic Claude structured output plans the replication, generates R or Python scripts, runs analyses, verifies results against published tables/figures within tight tolerances, and generates discrepancy reports. It's built around claude structured input for tasks like note taking and output json, turning vague replication goals into structured R outputs.

Why is it gaining traction?

It stands out with specialized agents—like epidemiology reviewers and R code checkers—that enforce quality gates (80+ scores to commit) and handle the full 6-phase pipeline from paper intake to polished reports. Developers dig the github llm structured output tools, claude structured output api integration, and hooks for session logging, making replications reproducible without manual drudgery. The contractor mode auto-plans, implements, and verifies, hooking users who hate chasing paper bugs.

Who should use this?

Epidemiologists or genetic researchers replicating UKB studies, like HR/OR estimates from cohort papers. Social scientists auditing causal claims in published work. R/Python users tired of manual data audits and script translation from Stata originals.

Verdict

Worth forking for UKB replication workflows—solid docs and claude structured output agent patterns make it instantly usable despite 15 stars and 1.0% credibility score. Still early and niche; test on one paper before scaling, as it demands Claude Code setup and your own data.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.