wanshuiyinCommunity·12 Agent Skills Included

Anti-Autoresearch

Forensic integrity audits of research papers and submissions

Audits research papers for self-contradictions, fabricated citations, missing baselines, invalid proofs, and evaluation design flaws. Builds a span-anchored evidence ledger so every finding quotes the exact sentence, file, and line it refers to. Replaces slow manual peer-review checking with a deterministic verdict and a reviewer-ready integrity report.

123|7

↗

npx skills add wanshuiyin/Anti-Autoresearch --all -g -y

All Skills in This Repository (12)

Pure Emerald Level Indicators

📦 In Repo

wanshuiyin

anti-autoresearch

Audit research papers for self-consistency and fabrication across multiple dimensions.

novelty-duplication-advisory

Audit paper submissions for trivial technique combinations and duplicate publications.

experiment-forensics

Audit experimental results against an evidence ledger using cross-model code auditing.

ai-style-impressions

Identify and annotate AI-origin writing patterns in documents.

consistency-audit

Audits PDF research papers for contradictions in method descriptions and numerical claims without external ground truth.

baseline-comparison-audit

Audit research paper baselines for completeness, fairness, and statistical significance.

adversarial-case-builder

Construct an adversarial case memo from evidence ledger and auditor findings.

eval-design-forensics

Audit research paper evaluation design and reporting validity for leakage and bias.

proof-derivation-forensics

Analyze LaTeX mathematical proofs for logical gaps and circular reasoning.

presentation-signals

Detect duplicate tables, leftover code strings, and unnecessary figures in research papers.

evidence-ledger

Construct deterministic evidence ledgers from LaTeX, PDF, and text papers.

citation-forensics

Validate citation integrity in academic papers against DBLP and arXiv sources.

Community

Advanced

Frequently Asked Questions

FAQPage Schema

How to install Anti-Autoresearch?▼

Run `npx skills add wanshuiyin/Anti-Autoresearch --all -g -y` in your terminal to install all skills and the orchestrating workflow globally.

How to check if a research paper is reliable?▼

Run the anti-autoresearch workflow on the paper's folder, PDF, or arXiv ID. It cross-checks numbers, citations, baselines, proofs, and evaluation design, then produces a reviewer-ready integrity report.

Can it detect fabricated or hallucinated citations?▼

Yes. The citation-forensics skill verifies every reference against DBLP, arXiv, and DOI records, flagging papers that do not exist, have wrong metadata, or are cited for claims they never make.

Does it work on PDF-only submissions without code?▼

Yes. It runs at observability level L0 for PDF-only papers, checking internal consistency and citations, and clearly marks which findings would need the code repository to confirm.

Is Anti-Autoresearch an AI-writing detector?▼

No. It never judges authorship or assigns AI probabilities. It only surfaces verifiable discrepancies in the paper's own evidence, with AI writing-style impressions listed separately at zero verdict weight.