paper-digest

Summarizes research paper PDFs into a structured nine-field digest with relevance scoring.

Updated Mar 21, 2026
One-click install
npx skills add https://github.com/pulak999/ai-tooling --skill paper-digest-pulak999
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: paper-digest
Source: https://github.com/pulak999/ai-tooling/tree/main/ai-assistant-kit/cursor-skills/paper-digest
Command: npx skills add https://github.com/pulak999/ai-tooling --skill paper-digest-pulak999

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires poppler, pymupdf.

What problem does it solve? Reading research papers is time-consuming, and abstracts rarely reveal the actual argument, limitations, or relevance to your work. This Skill extracts the core problem, hardness, baseline, solution, results, and technologies from paper PDFs into a consistent structured digest. ## Core Features & Use Cases - Structured 9-Field Digest: Produces PROBLEM, HARDNESS, BASELINE, SOLUTION, RESULTS, TECH, DELTA, RELEVANCE, and MY TAKE fields for every paper. - Relevance Scoring: Classifies each paper as new problem, new solution, or variant, and scores relevance 0-3 against your active research projects. - Batch Mode: Digests multiple PDFs at once and appends a batch summary listing must-read papers, new problems, and common technologies. - Use Case: You have a reading list of ten systems papers. Run a batch digest to get one comparable block per paper, identify the two must-reads, and spot shared techniques across the set. ## Quick Start Digest the attached paper PDF into the structured format and score its relevance to my current research.

Frequently Asked Questions about paper-digest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I summarize a research paper PDF into structured notes?

Provide the paper PDF and the digest is produced in a fixed nine-field format: PROBLEM, HARDNESS, BASELINE, SOLUTION, RESULTS, TECH, DELTA, RELEVANCE, and MY TAKE. Text is extracted with Poppler or PyMuPDF, and key sections are read in a targeted order.

How to summarize multiple research papers at once?

Supply multiple PDFs and one digest block is produced per paper, separated by dividers. A batch summary is appended listing must-read papers, directly relevant papers, new problems identified, and technologies common across the set.

What tools are needed to extract text from PDF papers?

The extraction script cascades through Poppler pdftotext for layout-preserving text, PyMuPDF via a virtualenv for resistant PDFs, and pdftoppm rasterization as a last resort. Run the pdf-summarizer install script to set up these dependencies.

Does PDF text extraction work on scanned or figure-heavy papers?

Text extraction is the primary method since academic papers are text-heavy. For pages where key figures or tables matter, individual pages are rasterized to JPEG at 150 DPI so they can be inspected visually.

How is paper relevance to my research scored?

Each paper receives a RELEVANCE score from 0 to 3 with a one-line justification, anchored to your known active projects. It also gets a DELTA classification marking it as a new problem, new solution, or a variant of known approaches.