literature-verifier

What problem does it solve?

This Skill verifies whether user-provided literature references are real and accurately reported, and it detects hallucinated or fabricated citations in both English and Chinese sources.

Core Features & Use Cases

• Reference authenticity verification: Checks existence and validates core metadata such as title, authors, venue/journal, year, DOI, ISBN, and URLs across multiple source types.
• Identifier-based workflows: Uses DOI resolution (CrossRef/doi.org), CrossRef title/author search, and URL reachability/metadata extraction; supports books via ISBN lookups.
• Chinese literature核查 (中文文献核查): Prioritizes CNKI direct search via Chrome CDP (MCP tools) and falls back to required multi-step WebSearch across CNKI/Wanfang/Baidu Scholar when CDP is unavailable.
• Hallucination pattern guidance: When verification fails, it recommends relevant hallucination patterns from provided reference docs to explain likely fabrication vs metadata errors.
• Batch reference auditing: Produces per-item verdicts and a summary table with confidence levels suitable for reference list cleanup.

Quick Start

Ask the AI to verify the reference list and return a table of confirmed vs likely real vs likely fabricated vs metadata-error results.