scientific-similarity-checkerlisted
Install: claude install-skill patolojiAI/pathology-skills-collection
# Scientific Similarity Checker
A multi-source literature search and academic integrity tool that analyzes a scientific article
against global databases to surface similar work and flag potential misconduct.
---
## Phase 1 — Extract Article Metadata
Before searching anything, extract as much metadata as possible from the input. The input may be:
- **PDF**: Use pdf-reading skill if available; otherwise read with bash `pdftotext` or similar
- **Image**: Read visually — extract title, authors, abstract from the image
- **Text/paste**: Parse directly
Extract these fields (mark each as `found` or `inferred` or `missing`):
```
title: (full title)
authors: (list: Last FM, Last FM, ...)
year: (publication or submission year)
journal: (journal/conference/preprint server name)
doi: (if present)
pmid: (if present)
abstract: (full abstract text)
keywords: (author-provided or extracted from abstract)
institution: (author affiliations if visible)
funding: (funding sources if visible)
corresponding: (corresponding author email/name)
```
For the abstract, if not explicitly labeled, look for the structured paragraph immediately after
author affiliations. Keywords may appear at the end of the abstract or in a separate section.
If the DOI is present, use it immediately to do a Crossref lookup (Phase 2a) — this often
resolves all metadata at once.
---
## Phase 2 — Multi-Database Search Strategy
Run these sear