技能详情（站内镜像，无评论）
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v1.0.3
统计：⭐ 2 · 673 · 7 current installs · 7 all-time installs
⭐ 2
安装量（当前） 7
🛡 VirusTotal ：良性 · OpenClaw ：良性
Package：aurora-1412/arxiv-reader
安全扫描（ClawHub）
- VirusTotal ：良性
- OpenClaw ：良性
OpenClaw 评估
This package appears to be what it claims: a Python-based arXiv paper reader that fetches papers, parses LaTeX, and uses an LLM to produce reading notes; no obvious malicious behavior is present, though there are a few minor metadata inconsistencies you should be aware of.
目的
The skill's code and SKILL.md match the description: it fetches arXiv papers, parses LaTeX, and calls an LLM to produce notes. However, the registry metadata lists only 'uv' as a required binary while the implementation clearly needs a Python runtime (it runs main.py and uses pip). Also the code reads ARXIV_* configuration values from .env (ARXIV_CATEGORIES, ARXIV_MAX_RESULTS, FETCH_LOOKBACK_DAYS) which are not listed in requires.env — these a…
说明范围
SKILL.md instructs creating a .env and running commands via the 'uv' wrapper; runtime instructions are limited to fetching/arXiv parsing and LLM calls. The code reads only project files (skills/, prompts/, .env) and fetches content from arXiv and the configured LLM endpoint; it does not access system-wide secrets or unrelated files. It will send paper text (including parsed LaTeX) to the LLM provider — expected for this functionality.
安装机制
There is no formal install spec in the registry (instruction-only install), but the repository includes requirements.txt and the SKILL.md tells users to run 'uv pip install -r requirements.txt'. That means dependencies will be installed from PyPI (moderate risk compared to no install spec). Also the requirements include an unversioned 'arxiv-to-prompt' entry — you may wish to inspect that package on PyPI before installing.
证书
The declared required env vars (LLM_API_KEY, LLM_BASE_URL, LLM_TEMPERATURE, LLM_MAX_TOKENS) are proportional to an LLM-based reader; LLM_API_KEY is the primary credential as expected. Two small notes: (1) config.py also reads ARXIV_* and FETCH_LOOKBACK_DAYS from .env but these are not listed in requires.env (they are not secrets but are configuration values). (2) The skill will transmit full paper text (and parsed LaTeX) to the configured LLM …
持久
The skill does not request permanent 'always' inclusion and does not change other skills or system-level settings. It runs as a normal user-level Python program and does not request elevated privileges.
综合结论
This skill appears coherent and implements the advertised functionality, but review the following before installing: - Secrets and data flow: you must provide LLM_API_KEY and LLM_BASE_URL — the tool will send full paper text (including LaTeX) to that LLM provider. Only use an API/key you trust and do not supply keys for services you do not want to see this data. - Binary requirement mismatch: registry lists 'uv' as required, but the code requi…
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「arxiv_deep_reader」。简介：利用python，指定某个arxiv_id/url， 基于 LLM Agent 对这篇arxiv论文进行分类与深度阅读，直接print打印阅读笔记。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/aurora-1412/arxiv-reader/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
---
name: Arxiv Paper Reader
description: 利用python，指定某个arxiv_id/url， 基于 LLM Agent 对这篇arxiv论文进行分类与深度阅读，直接print打印阅读笔记
metadata: {"openclaw":{"requires":{"bins":["uv"],"env":["LLM_API_KEY","LLM_BASE_URL","LLM_TEMPERATURE","LLM_MAX_TOKENS"]},"primaryEnv":"LLM_API_KEY"}}
---
## 快速开始
### 1. 配置 `.env`
```bash
cp .env.example .env # 或直接编辑 .env
```
确定你已经配置了：
- `LLM_API_KEY` — OpenAI 或兼容 API 的密钥
- `LLM_BASE_URL` — API 地址
### 2. 运行
```bash
uv venv
uv pip install -r "{baseDir}/requirements.txt"
# 单篇论文模式：指定 arxiv_id 或 URL
uv run python "{baseDir}/main.py" --arxiv-id 2401.12345
uv run python "{baseDir}/main.py" --arxiv-id https://arxiv.org/abs/2401.12345
uv run python "{baseDir}/main.py" --arxiv-id https://arxiv.org/pdf/2401.12345.pdf
# 指定以特定类别阅读
uv run python "{baseDir}/main.py" --arxiv-id xxxx --category yyy
# 查看所有类别
uv run python "{baseDir}/main.py" --list
```
## 添加新的阅读分类
在 `skills/` 下新建文件夹，包含两个文件：
```
skills/your_new_category/
├── _metadata.md # 分类描述（告诉 Classifier 什么论文属于这个类别）
└── reading_prompt.md # 阅读指南（告诉 Reader Agent 重点关注什么）
```
重启即可自动识别，无需修改任何代码。
## Python包
- LangChain 1.x — Agent 框架（基于 LangGraph）
- LangChain OpenAI — LLM 接口（兼容 DeepSeek 等 OpenAI-compatible API）
- arxiv — 官方 Python 库
- arxiv-to-prompt 获取arxiv论文latex源码