paper-claim-audit

Extract quantitative claims from papers and verify them against raw results.

Updated Jul 2, 2026
One-click install
npx skills add https://github.com/SamyakJhaveri/loam --skill paper-claim-audit-samyakjhaveri
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: paper-claim-audit
Source: https://github.com/SamyakJhaveri/loam/tree/main/seed/_research/skills/paper-claim-audit
Command: npx skills add https://github.com/SamyakJhaveri/loam --skill paper-claim-audit-samyakjhaveri

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Zero-context verification ensures that every numeric claim in a manuscript is faithful to the underlying data, preventing confirmation bias and data misreporting.

Core Features & Use Cases

  • Zero-context workflow: The auditor accesses only the paper claims and raw results, avoiding prior assumptions.
  • Traceability: Each claim is linked to exact evidence in the data files with a clear status (exact_match, rounding_ok, mismatch).
  • Reproducibility support: Generates machine-readable artifacts (PAPER_CLAIM_AUDIT.json) for downstream checks and audits.

Quick Start

Run the audit against the paper's source and its raw results to produce a structured report.

Frequently Asked Questions about paper-claim-audit

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

FAQPage Schema
How do I verify quantitative claims in a research paper against raw data?

Paper claim auditing identifies every quantitative claim in a manuscript and extracts the claimed values. The audit matches these claims against exact data evidence from raw results to produce a report with per-claim statuses and a final verdict.

What is zero-context verification for manuscript data reproducibility?

Zero-context verification is an auditing approach where the tool accesses only the manuscript claims and raw results without prior assumptions. It ensures every numeric claim is faithful to the underlying data, preventing confirmation bias and data misreporting.

How do I generate machine-readable artifacts for paper claim audits?

You generate machine-readable artifacts by running the audit against the paper source and raw results. The process outputs a PAPER_CLAIM_AUDIT.json file containing structured per-claim statuses for downstream reproducibility checks and audits.

How does traceability work when checking paper claims against data evidence?

Traceability works by linking each extracted claim to exact evidence within the raw data files. The audit assigns a clear status—such as exact_match, rounding_ok, or mismatch—to every claim, ensuring fidelity in the verification report.

Can I audit research papers without prior knowledge of the study's assumptions?

Yes, the zero-context workflow is designed to audit papers without prior knowledge of the study's assumptions. It strictly isolates the paper claims and raw results to verify quantitative data without introducing external confirmation bias.