read-arxiv-paper

What problem does it solve?

This Skill enables researchers to read and summarize arXiv papers by programmatically fetching the LaTeX source (not the PDF), unpacking it locally, identifying a primary entrypoint (e.g., main.tex), and generating a concise summary that integrates with the nanochat knowledge store.

Core Features & Use Cases

• URL normalization to arXiv source URLs and subsequent access to LaTeX sources.
• Local caching of the LaTeX tarball under ~/.cache/nanochat/knowledge/{arxiv_id} for quick re-reads.
• Unpacking & entrypoint discovery to locate main.tex or equivalent starting point.
• Recursive source reading to build a coherent, in-depth understanding of the paper.
• Knowledge integration by writing a Markdown summary to ./knowledge/summary_{tag}.md and relating it to nanochat context.

Quick Start

Use the read-arxiv-paper skill with the arXiv URL https://arxiv.org/abs/2601.07372 to generate a knowledge summary.