Agent Skills › Prismer-AI/Prismer › arxiv-reader

arxiv-reader

GitHub

用于读取和分析arXiv学术论文，支持获取摘要、章节结构及完整LaTeX文本，便于LLM进行内容研究。

skills/arxiv-reader/SKILL.md Prismer-AI/Prismer

Trigger Scenarios

用户提及arXiv、预印本或学术文献提供论文ID并要求阅读或分析

Install

npx skills add Prismer-AI/Prismer --skill arxiv-reader -g -y

More Options

Use without installing

npx skills use Prismer-AI/Prismer@arxiv-reader

指定 Agent (Claude Code)

npx skills add Prismer-AI/Prismer --skill arxiv-reader -a claude-code -g -y

安装 repo 全部 skill

npx skills add Prismer-AI/Prismer --all -g -y

预览 repo 内 skill

npx skills add Prismer-AI/Prismer --list

SKILL.md

Frontmatter

{
    "name": "arxiv-reader",
    "metadata": {
        "openclaw": {
            "tags": [
                "arxiv",
                "research",
                "academic",
                "papers",
                "latex"
            ],
            "emoji": "📄",
            "requires": {
                "os": [
                    "darwin",
                    "linux",
                    "win32"
                ],
                "bins": []
            }
        }
    },
    "description": "Read and analyze arXiv papers by fetching LaTeX source, listing sections, or extracting abstracts. Use when the user mentions arXiv, research papers, preprints, paper IDs like 2301.xxxxx, or wants to read academic publications."
}

arxiv-reader

Read and analyze arXiv papers directly from the workspace. Converts LaTeX source into clean text suitable for LLM analysis.

Description

Fetches arXiv papers, flattens LaTeX includes, and returns clean text for LLM analysis. Works standalone (downloads directly from arXiv) or delegates to the container arXiv server when available. Results are cached locally for instant repeat access.

Usage Examples

"Read the paper 2301.00001 from arXiv"
"What sections does paper 2405.12345 have?"
"Get the abstract of 2312.09876"
"Fetch paper 2301.00001 without the appendix"

Process

Quick look — Use arxiv_abstract to get a paper's abstract before committing to a full read
Survey structure — Use arxiv_sections to understand the paper's outline
Deep read — Use arxiv_fetch to get the full flattened LaTeX for analysis

Tools

arxiv_fetch

Fetch the full flattened LaTeX source of an arXiv paper.

Parameters:

arxiv_id (string, required): arXiv paper ID (e.g. 2301.00001 or 2301.00001v2)
remove_comments (boolean, optional): Strip LaTeX comments (default: true)
remove_appendix (boolean, optional): Remove appendix sections (default: false)
figure_paths (boolean, optional): Replace figures with file paths only (default: false)

Returns: { content: string, arxiv_id: string, cached: boolean }

Example:

{ "arxiv_id": "2301.00001", "remove_appendix": true }

arxiv_sections

List all sections and subsections of an arXiv paper.

Parameters:

arxiv_id (string, required): arXiv paper ID

Returns: { arxiv_id: string, sections: string[] }

Example:

{ "arxiv_id": "2301.00001" }

arxiv_abstract

Extract just the abstract from an arXiv paper.

Parameters:

arxiv_id (string, required): arXiv paper ID

Returns: { arxiv_id: string, abstract: string }

Example:

{ "arxiv_id": "2301.00001" }

Notes

All results are cached locally — repeat requests are instant
Works standalone (no Docker required) or with the container arXiv server
Paper IDs support version suffixes (e.g. 2301.00001v2)
Very large papers may take 10-30 seconds on first fetch
arxiv_abstract uses the arXiv Atom API for fast metadata retrieval in standalone mode