A specialized tool for validating academic citations and detecting AI-generated hallucinations across global research databases.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install paper-reference-checker
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install paper-reference-checker using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Paper Reference Checker?
The Paper Reference Checker is a robust utility within the Openclaw Skills framework designed to ensure the integrity of academic writing. By systematically cross-referencing bibliographies against authoritative sources like Google Scholar, arXiv, and CNKI, it identifies fabricated or non-existent citations often produced by large language models. This skill provides researchers and developers with a reliable way to audit papers, verify DOIs, and maintain high standards of academic rigor.
Built to be token-efficient, this tool handles various input types ranging from raw PDF uploads to direct Overleaf and arXiv links. It serves as a critical layer of defense against AI hallucinations, providing peace of mind for anyone using automated agents to assist in research or literature reviews through Openclaw Skills.
Paper Reference Checker Use Cases
- Verifying citations in research papers to ensure all sources genuinely exist.
- Detecting AI-hallucinated references in drafts generated by language models.
- Validating DOIs and arXiv IDs in bibliographies before submission.
- Auditing academic documents for broken or uncertain citations during peer review.
- Cleaning up BibTeX files by identifying unresolvable entries.
How Paper Reference Checker Works
- Extract citations from the source document using token-efficient methods, such as targeting the final pages of a PDF or regex-parsing .tex files.
- Resolve primary identifiers like DOIs and arXiv IDs to confirm existence through authoritative resolution services.
- Query global academic databases including Google Scholar, Semantic Scholar, and CNKI for remaining titles.
- Categorize each reference with authenticity labels such as Verified, Uncertain, or Fabricated based on query results.
- Output a structured report detailing the verification status of every reference found in the document.
Paper Reference Checker Setup
To integrate this capability into your workflow, ensure your environment is configured for Openclaw Skills. Use the following commands to get started:
# Install the Paper Reference Checker skill
openclaw add paper-reference-checker
# Configure access for specialized databases if necessary
openclaw config paper-reference-checker --set-api-keys
Paper Reference Checker Data Schema & Taxonomy
The Paper Reference Checker organizes its verification logic and results using the following structure:
| Component | Description |
|---|---|
references/ |
Directory containing rules for extraction and search tactics. |
scripts/ |
Contains the core decision tree logic for authenticity judgment. |
examples/ |
Provides sample BibTeX files and annotated verification reports. |
Citations are processed into a metadata taxonomy including labels like VERIFIED (confirmed in ≥1 DB), NOT FOUND (no matches across channels), and FABRICATED (non-existent venue or unresolvable DOI).
Paper Reference Checker Advanced Features
- Token-efficient extraction focusing on the last 15-20% of PDF pages to reduce overhead.
- Multi-platform querying support across PubMed, IEEE Xplore, ACM DL, and DBLP.
- Seamless integration with Overleaf projects via .tex and .bib file parsing.
- Automated DOI resolution and Crossref metadata validation.
- Support for Chinese academic databases like CNKI for comprehensive global verification within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Buddy: ASCII Pet Companion
Buddy is an interactive ASCII pet companion named Flicker that provides personality-driven commentary and snarky feedback within your AI agent workspace.
Baidu Search
A high-performance web search skill leveraging the Baidu AI Search Engine to provide real-time information and research capabilities to AI agents.
AI-First Engineering
An engineering operating model designed to optimize team workflows, architecture, and testing standards for AI-assisted code generation.
Chinese Invention Patent Drafting Assistant
A specialized AI assistant that guides users through a multi-step dialogue to draft professional Chinese invention patent documents including claims and descriptions.
di-perplexity-style
An advanced research and analysis skill that implements a Perplexity-style multi-step methodology for deep investigation and actionable insights.
Apollo Circadian
Apollo Circadian is an intelligent scheduling skill that synchronizes AI task execution with human biological rhythms to maximize cognitive efficiency.
Featured*
