audit-reproducibility

Verify numeric claims in academic manuscripts against analysis outputs with tolerance thresholds.

Updated Jun 27, 2026

Search Tags:#academic research #reproducibility #manuscript analysis #numeric verification #tolerance thresholds

Ready to run

npx skills add https://github.com/fredmilhome/laffer_tobacco --skill audit-reproducibility-fredmilhome

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: audit-reproducibility
Source: https://github.com/fredmilhome/laffer_tobacco/tree/main/.claude/skills/audit-reproducibility
Command: npx skills add https://github.com/fredmilhome/laffer_tobacco --skill audit-reproducibility-fredmilhome

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill addresses the challenge of ensuring numerical reproducibility in academic manuscripts by comparing numeric claims in a manuscript against actual analysis outputs.

Core Features & Use Cases

Numeric Claim Verification: Cross-checks numeric claims (estimates, p-values, counts) in a manuscript against analysis outputs.
Tolerance Thresholds: Applies tolerance thresholds to verify compliance with replication protocols.
Before Submission: Use before submission to catch discrepancies that could affect publication.
Before Replication Package Release: Ensures the code reproduces the paper as released.

Quick Start

Run the 'audit-reproducibility' skill with the manuscript path and output directory: audit-reproducibility manuscript_path outputs_directory

Frequently Asked Questions about audit-reproducibility

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

FAQPage Schema

How do I verify numeric claims in an academic manuscript against analysis outputs?▼

Verifying numeric claims in an academic manuscript involves cross-checking reported estimates, p-values, and counts against actual analysis outputs by applying predefined tolerance thresholds to ensure reproducibility.

What is numeric reproducibility and when do I need to check it?▼

Numeric reproducibility ensures that the numbers reported in an academic manuscript match the results generated by the analysis code. You need to check it before journal submission or releasing a replication package.

How do I check manuscript reproducibility before releasing a replication package?▼

Checking manuscript reproducibility before releasing a replication package requires running an audit that compares the manuscript's reported values against computed results using predefined tolerance thresholds.

Can I apply custom tolerance thresholds when verifying p-values and estimates?▼

Applying custom tolerance thresholds when verifying p-values and estimates allows you to define acceptable variance limits between manuscript claims and analysis outputs to meet specific replication protocol compliance.

What is the best way to catch numeric discrepancies in a research paper?▼

The best way to catch numeric discrepancies in a research paper is to automate the comparison of manuscript text against analysis outputs, applying tolerance thresholds to flag values that fail replication compliance.

Does manuscript numeric verification work without specific analysis output files?▼

Manuscript numeric verification requires specific analysis output files to compare against the manuscript's reported values, as the core task depends on contrasting claimed numbers with computed results.