Paper Integrity Forensics
Paste the following prompt into your AI chat to install this skill:
Follow https://skillhub.cn/install/skillhub.md to install @user_164f4c1f/paper-integrity-forensics.
About this skill
Problem
When citing a paper, the abstract often hides serious risks: retraction, predatory or hijacked journals, inconsistent statistics, image reuse, fabricated references, and mismatched author affiliations. Treating a weak load-bearing reference as a foundation can contaminate downstream research. This skill frames pre-citation review as a reproducible evidence workflow rather than a simple yes/no judgment.
How It Works
Input can be a title, DOI, PDF, reference entry, or database link. The workflow runs through seven dimensions:
- Identity and existence: checks DOI resolution, journal name, issue, and page ranges against Crossref, PubMed, CNKI, or similar indexes.
- Retraction and integrity records: uses the bundled free retraction API to query aggregated Retraction Watch and PubPeer signals, including
retracted,concern,corrected, and peer comments. - Journal legitimacy: combines DOAJ, Cabells, and index coverage to assess predatory or hijacked-journal risk.
- Statistical and image integrity: flags possible
p-hacking, GRIM/statcheck inconsistencies, reused blots, or AI-generated figures, with guidance to external tools. - Citation-chain audit: for papers with self-citation >30% or heavy concentration in one journal/author group, extracts load-bearing references for batch verification.
The output is usually a self-contained HTML report with four-level risk ratings, evidence links, red flags, and a citation recommendation.
Boundaries
It aggregates evidence and provides risk screening, but it does not make final research-misconduct determinations, automatically recompute raw statistics, build private image databases, or exhaustively query journal blacklists. When DOI, full text, or images are unavailable, the relevant dimensions are marked as insufficient and a follow-up checklist is provided.
Use Cases
- Check whether a reference is retracted or has peer challenges before citing it in a manuscript.
- Assess journal legitimacy and reference-chain authenticity when building a review article argument.
- Screen statistical, image, and author-affiliation red flags and generate an evidence-linked risk report.
- Verify Chinese papers with CNKI, Wanfang, and retraction notices when DOI data is missing.
Best For
- Graduate students who need to screen thesis references for retractions, fabricated citations, and image-reuse risks.
- Faculty reviewers who need to judge whether cited literature is reliable before advising revisions.
- Researchers who need to verify core supporting papers’ journal legitimacy, statistical flags, and citation chains before submission.
- Academic editors who need to check reference lists for retraction notices, expressions of concern, and peer challenges.
Related Skills
Based on Hengsheng Juyuan MCP financial data, it assesses equity, bond, commodity, and overseas markets and produces sourced risk reports with allocation guidance.
Applies a six-question industry-chain framework for structured fundamental analysis of sectors, profit pools, competition, ROIC, valuation, and domestic substitution.
Performs structured contract risk assessment, adds revision suggestions in the document, and generates a standalone review report.
A single-stock research framework that produces seven-dimension drafts, integrated analysis, scenario projections, and dual-track reference, with optional comparison and no investment advice.