Pdf classifier

Skill cxcscmu/SkillLearnBench/skills/b4-skill-creator-claude-opus-4-6/organize-messy-files/pdf-classifier

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install

npx -y skills add cxcscmu/SkillLearnBench --skill pdf-classifier

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Classifies PDF academic papers by subject area using text extraction from first pages. Use this skill whenever you need to categorize, sort, or organize PDF papers by topic, research field, or subject matter.

SKILL.md

2.8 KB, as published. Nobody here has run it

PDF Classifier

Extracts text from PDF files and classifies them into predefined subject categories based on title, abstract, and keywords.

Approach

Extract text from the first 1-2 pages of each PDF using Python's PyPDF2 or pdfplumber library
Classify based on keyword matching against subject-specific vocabulary
Handle ambiguity by scoring each category and picking the highest match

Classification Strategy

For each document, extract text and look for domain-specific keywords:

Keyword Sets by Subject

LLM: language model, transformer, attention mechanism, GPT, BERT, token, prompt, fine-tuning, RLHF, LLM, neural network language, text generation, NLP, natural language processing, large language, instruction tuning, in-context learning, chain-of-thought, retrieval-augmented, embedding, tokenizer, decoder, encoder
Trapped ion & quantum computing: trapped ion, quantum computing, quantum gate, qubit, quantum error correction, ion trap, quantum entanglement, quantum algorithm, quantum circuit, quantum processor, quantum simulation, quantum information, Paul trap, Coulomb crystal, motional mode, Raman transition, quantum logic, fault-tolerant quantum, surface code, topological quantum
Black hole: black hole, event horizon, Hawking radiation, singularity, gravitational wave, general relativity, spacetime, Schwarzschild, Kerr metric, accretion disk, gravitational lensing, binary merger, LIGO, dark matter, cosmological, neutron star merger, gravitational collapse, Penrose, ergosphere, information paradox
DNA: DNA, genomic, gene expression, nucleotide, genome, CRISPR, sequencing, transcription, chromosome, mutation, protein folding, epigenetic, RNA, polymerase, genetic, double helix, base pair, replication, molecular biology, bioinformatics
Music history: music history, musical, composer, symphony, opera, baroque, classical period, romantic era, jazz history, musicology, ethnomusicology, sonata, concerto, counterpoint, harmony theory, musical form, orchestration, medieval music, renaissance music, blues history

Implementation

Use Python with PyPDF2 for text extraction. For each file:

Open PDF reader
Extract text from pages 0-1
Lowercase the text
Count keyword hits per category
Assign to category with most hits (default to music_history if no clear match as the catch-all)

Edge Cases

If text extraction fails (scanned PDFs), use the arXiv ID in the filename to infer subject via the ID prefix pattern
arXiv categories: astro-ph, gr-qc, hep-th → black_hole; quant-ph → trapped_ion_and_qc; cs.CL, cs.AI, cs.LG → LLM; q-bio → DNA