Source: https://github.com/aipoch/medical-research-skills

When to Use

You have a batch of references and need a publication year distribution table (counts and percentages).
You need a journal distribution table (Top N optional) for a literature review or report appendix.
Your input is pasted citations (BibTeX/RIS/EndNote/plain text/mixed) and you want quick aggregation.
Your input is local reference files (.bib/.ris/.txt/.csv) and you want consistent, standardized output.
You have a local PDF folder and want to extract year/journal signals (best-effort) and summarize them.

Key Features

Supports multiple input types: pasted text, local reference files, and local PDF directories (via script).
Extracts Year and Journal using format-specific parsing rules (BibTeX/RIS/plain text/PDF).
Produces two standardized tables:
- Year distribution: year, count, percent
- Journal distribution: journal title, count, percent
Provides a summary including totals and unknown-field counts (unknown year / unknown journal).
Conservative extraction: does not guess when metadata is unclear; ambiguous items are counted as unknown.
Local-only operation: no network calls, no external APIs, no credential usage.