Scientific Fraud Detection MCP

MCP server exposing 8 scientific-integrity tools that screen research literature for fraud, misconduct, and questionable research practices. Powered by 16 Apify actors spanning academic databases, biomedical repositories, patent records, and archival sources.

Tools

#	Tool	Method	Price
1	audit_statistical_consistency	SPRITE/GRIM tests + Benford chi-squared	$0.040
2	analyze_p_curve_z_curve	P-curve right-skew/flatness + z-curve EM mixture	$0.035
3	fit_selection_model_meta_analysis	Vevea-Hedges selection model + DerSimonian-Laird RE	$0.045
4	detect_citation_network_anomalies	TERGM citation rings + self-citation Gini	$0.040
5	screen_image_data_forensics	Benford + DCT anomaly + MinHash duplicate detection	$0.045
6	trace_causal_contamination	BFS + do-calculus + Dirichlet process clustering	$0.040
7	detect_harking_bayesian_surprise	KL(posterior‖prior) + Brier score calibration	$0.035
8	self_calibrate_detection	Platt scaling fixed-point loop (Godelian self-audit)	$0.040

Data Sources (16 Actors)

Academic: OpenAlex, Crossref, CORE, DBLP, ArXiv Biomedical: PubMed, Semantic Scholar, Europe PMC, ORCID Clinical: NIH Grants, Clinical Trials Archival: Wayback Machine, Website Content Technical: GitHub, USPTO Patents, Hacker News

Quick Start

npm install
npm run build
npm start          # port 3024

Connect via SSE (GET /sse) or Streamable HTTP (POST /mcp).