This tool serves as a live Model Context Protocol (MCP) server designed to automate the rigorous auditing of scientific literature for various forms of fraud and questionable research practices. It orchestrates parallel queries across numerous academic databases to apply 11 forensic algorithms, including GRIM/SPRITE statistical auditing, p-curve analysis, Vevea-Hedges selection models, TERGM citation anomaly detection, Benford DCT forensics, do-calculus contamination tracing, and Bayesian surprise HARKing detection. Built for research integrity officers, meta-analysts, and AI assistants, it provides structured JSON output for immediate downstream analysis, significantly reducing the manual effort and time required for comprehensive integrity assessments.
주요 기능
010 GitHub stars
02GRIM, SPRITE, and Benford's law for statistical consistency auditing
03Benford DCT and MinHash LSH for image manipulation and text duplication screening
04P-curve and z-curve analysis for p-hacking detection and replicability estimation
05TERGM and Gini coefficient for detecting citation rings and self-citation anomalies
06Vevea-Hedges selection model for publication-bias-corrected meta-analysis
사용 사례
01Replication crisis meta-analysis to estimate Expected Replication Rate and detect systematic p-hacking
02Research integrity assessment for universities, funding agencies, and journals
03Systematic review and meta-analysis support to estimate publication-bias-corrected effects and flag problematic sources
