paper-reader

Generates structured Chinese lecture-style reading notes from research papers into NOTES.md.

609|50|Updated Jul 7, 2026
One-click install
npx skills add https://github.com/poco-ai/Agentero --skill paper-reader-poco-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: paper-reader
Source: https://github.com/poco-ai/Agentero/tree/main/templates/vault/.agents/skills/paper-reader
Command: npx skills add https://github.com/poco-ai/Agentero --skill paper-reader-poco-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Reading dense research papers is slow and fragmented: PDFs, highlights, and notes live in different places, and it is hard to produce consistent, well-structured study notes. This Skill turns a paper in your local vault into a rigorous, structured Chinese reading report with figures, formulas, and citations. ## Core Features & Use Cases - Structured Paper Analysis: Produces a fixed-structure report covering problem and motivation, core contribution deconstruction, experiments and SOTA comparison, plus limitations and future work. - Figure-Aware Explanation: Extracts key figures from the paper source (converting PDF figures with pdftoppm when needed) and embeds them inline alongside the explanation. - Vault-Native Output: Writes lecture-style notes with frontmatter, Obsidian wikilinks, and local PDF citations into {paper}/NOTES.md, then marks the paper as read via the agentero CLI. - Author Lookup Support: Includes a reference guide for finding first/corresponding author emails, homepages, GitHub, ORCID, and OpenReview links via OpenAlex, DBLP, and other APIs. - Use Case: Import an arXiv paper into your Agentero vault, then ask the agent to read it; you get a complete Chinese NOTES.md with figures, LaTeX formulas, and linked citations ready for review. ## Quick Start Read the paper in papers/1706.03762 and write a structured Chinese lecture-style note with figures and citations into its NOTES.md.

Frequently Asked Questions about paper-reader

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

FAQPage Schema
How do I generate structured reading notes from a research paper?

Point the agent at a paper folder in your vault and it produces a fixed-structure report covering problem, motivation, method deconstruction, experiments, and limitations. The output is written as lecture-style Chinese notes into the paper's NOTES.md file.

How do I extract figures from a paper PDF for my notes?

Figures referenced from TeX sources are embedded directly with Markdown image syntax. For PDF-format figures, the workflow converts them with pdftoppm -jpeg -r 200 -singlefile and references the resulting .jpg file in the notes.

How do I find an author's email, homepage, or OpenReview profile?

Start with the paper's PDF or TeX author block for emails, then use the OpenAlex API to map the title to author IDs, ORCID, and affiliations. DBLP and GitHub search help confirm homepages and accounts, while OpenReview pages are best found via web search.

Does this work with Obsidian wikilinks and local vaults?

Yes, notes preserve and create Obsidian-style wikilinks using canonical vault-relative paths, and citations link directly to local PDF sections, figures, and pages. Existing user wikilinks are kept intact and only verified targets are linked.

What happens when a paper does not specify a detail like hyperparameters?

The workflow explicitly forbids fabricating missing details. Any information not stated in the paper, such as training hyperparameters or hardware, is marked as not explicitly specified in text rather than inferred.