scientific-integrity

Multi-gate AI research verification system that audits hypotheses, citations, experiments, results, and logic fallacies in AI-generated research to prevent hallucinations, citation fabrication, and methodological errors

Governance framework defining institutional authority boundaries and policy contracts for AI-assisted workflows.

Scientific Fraud Detection MCP Server

Heuristic, phase-specific screening for 2D MX2 semiconductor candidates, published with its full critical-review trail. Independent chemist audit verdict: "not scientifically reliable in its current form". A high-school exploratory project, not experimentally validated - do not use to guide synthesis, safety, investment, or engineering decisions.

An experimental audit engine that treats missing, conflicting, malformed, or unverified evidence as unresolved—not passed—with exact structural checks, adversarial false-claim tests, and proof-tool receipts that never count as proof on their own.

A gym for scientific judgment. AI agents render fatally flawed analyses at publication quality — we measured whether anyone (human or model) can tell. Live: https://confounded.vercel.app

开源论文体检工具 PaperGuard：扫描论文 PDF 的统计自洽、数字一致与图片完整性疑点，帮助科研人员发现学术不端线索，MIT 开源。Paper screening for scientific integrity.

Screen scholarly PDFs and Source Data for integrity red flags—image reuse, table anomalies, metadata—offline CLI and MCP. No API key required.

OpenTimestamps cryptographic proofs for Sprinkling Act published reports