mj-paper-flow

What problem does it solve?

Processing academic papers traditionally requires manually switching between tools to read content, write structured analysis documents, and create visual summary cards, a disjointed and time-consuming process. This skill automates the entire end-to-end workflow to eliminate that friction.

Core Features & Use Cases

• End-to-end paper processing: Accepts arxiv links, PDF files, or paper names as input, and automatically generates structured Lark Doc analysis and visual sketchnote PNG cards for each paper.
• Cost-optimized model routing: Uses Sonnet for paper reading tasks to reduce inference costs, and reserves Opus for visual card creation to maintain high aesthetic quality.
• Parallel multi-paper support: Processes multiple papers in parallel for paper reading, then handles card generation and doc embedding sequentially per paper to balance speed and quality.
• Use case: Researchers, students, or knowledge workers who need to process batches of academic papers can get both detailed written analysis and shareable visual cards in one go, without manual tool switching or context switching.

Quick Start

Provide one or more arxiv links, PDF files, or paper names to the mj-paper-flow skill to receive both structured Lark Doc analysis and visual summary cards for each input paper.