Verify numerical claims in research papers against raw evidence.
npx skills add https://github.com/jkfee/Auto-Research --skill paper-claim-audit
Please help me install this Agent Skill. Skill: paper-claim-audit Source: https://github.com/jkfee/Auto-Research/tree/main/skills/paper-claim-audit Command: npx skills add https://github.com/jkfee/Auto-Research --skill paper-claim-audit
This Skill provides zero-context verification that every numerical claim in a research paper matches the raw data, reducing confirmation bias and increasing trust in published results.
Provide the paper source and results artifacts to initiate a zero-context audit and generate a structured, machine-readable report.
High-intent search queries and answers about installing and using this skill.
Zero-context auditing prevents prior-context bias by using a fresh reviewer thread to verify claims. This process ensures end-to-end verification of numerical claims against raw evidence before paper submission, reducing confirmation bias.
You detect rounding drift and cherry-picking by applying a zero-context audit across paper sections, tables, and figures. This process enforces end-to-end verification by mapping every numerical claim directly to its raw evidence.
Zero-context verification is an auditing process that uses a fresh reviewer thread to check numerical claims against raw data without prior-context bias. It ensures full traceability of claims to evidence before submission.
Yes, you can audit figures and tables to find scope overclaims by cross-checking all numerical claims against raw results files. The audit applies across sections, tables, and figures to ensure claims match raw evidence.
To start a paper claim audit, you need to provide the paper source and results artifacts. This initiates the zero-context verification process and generates a structured, machine-readable report tracing claims to evidence.
anilcan-karaExtract text from PDFs and scanned documents using pymupdf or marker-pdf.
DarkbluelrConvert Epic requirements into a validated Slice DAG and generate a machine-readable Knife Plan.
dogekikiAudit experimental integrity by cross-referencing scripts, results, and claims against fraud patterns.
HolobiomicsLabEvaluate scientific claims by analyzing methodology, biases, and evidence strength.
