paper-claim-audit

To verify research paper data, you can perform a zero-context audit that compares every numerical claim in your .tex files against raw result files to eliminate confirmation bias and ensure accurate data reporting.

Zero-context claim verification means validating paper claims against raw data with no prior context, providing objective data integrity checks that eliminate confirmation bias before research paper submission.

You need the mcp__codex__codex tool with the gpt-5.5 model, Bash, Read, Write, Edit, Grep, and Glob to execute the paper claim audit and validate research integrity.

To audit specific numerical claims, provide the paper directory containing your .tex and raw result files, and the skill will perform customizable analysis to validate reported results match the raw evidence.

The best way to check research integrity is running a zero-context audit to ensure all reported results, numerical claims, and scope claims are accurately represented and match the raw evidence before submission.