← 返回 Skills 市场
Arxiv Watcher
作者
constx1337
· GitHub ↗
· v1.0.0
· MIT-0
281
总下载
1
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install arxiv-watch
功能描述
Fetch and monitor latest arXiv papers by category, generate summaries, and manage starred papers for easy reference.
使用说明 (SKILL.md)
\r \r
ArXiv Watcher\r
\r Monitor and track new papers on arXiv.org by category.\r \r
Overview\r
\r This skill helps you:\r
- Fetch latest papers from specific arXiv categories\r
- Generate formatted Markdown summaries\r
- Star/bookmark papers for later reading\r \r
Quick Start\r
\r
# Check latest papers in a category\r
/arxiv-watcher cs.AI\r
\r
# Multiple categories\r
/arxiv-watcher cs.AI cs.CL cs.LG\r
\r
# View starred papers\r
/arxiv-watcher --starred\r
```\r
\r
## Available Categories\r
\r
### Computer Science (cs.*)\r
| Category | Description |\r
|----------|-------------|\r
| cs.AI | Artificial Intelligence |\r
| cs.CL | Computation and Language |\r
| cs.CV | Computer Vision |\r
| cs.LG | Machine Learning |\r
| cs.NE | Neural and Evolutionary Computing |\r
| cs.RO | Robotics |\r
| cs.SE | Software Engineering |\r
\r
### Physics (physics.*)\r
| Category | Description |\r
|----------|-------------|\r
| astro-ph | Astrophysics |\r
| cond-mat | Condensed Matter |\r
| hep-ex | High Energy Physics - Experiment |\r
| hep-th | High Energy Physics - Theory |\r
| quant-ph | Quantum Physics |\r
\r
### Mathematics (math.*)\r
| Category | Description |\r
|----------|-------------|\r
| math.AG | Algebraic Geometry |\r
| math.CO | Combinatorics |\r
| math.DG | Differential Geometry |\r
| math.NT | Number Theory |\r
| math.ST | Statistics Theory |\r
\r
### Statistics (stat.*)\r
| Category | Description |\r
|----------|-------------|\r
| stat.ML | Machine Learning |\r
| stat.TH | Statistics Theory |\r
\r
Full category list: https://arxiv.org/category_taxonomy\r
\r
## Commands\r
\r
### Fetch Latest Papers\r
\r
```bash\r
python scripts/arxiv_watcher.py fetch \x3Ccategory> [--limit N]\r
```\r
\r
Fetches recent papers from arXiv. Outputs Markdown format with:\r
- Paper title and arXiv ID\r
- Authors\r
- Abstract preview (first 200 chars)\r
- Link to full paper\r
\r
### Star Papers\r
\r
```bash\r
python scripts/arxiv_watcher.py star \x3Carxiv_id>\r
```\r
\r
Adds a paper to starred list for later reference.\r
\r
### Unstar Papers\r
\r
```bash\r
python scripts/arxiv_watcher.py unstar \x3Carxiv_id>\r
```\r
\r
Removes a paper from starred list.\r
\r
### View Starred Papers\r
\r
```bash\r
python scripts/arxiv_watcher.py starred\r
```\r
\r
Lists all starred papers with their details.\r
\r
## Output Format\r
\r
Papers are displayed in Markdown format:\r
\r
```markdown\r
## [2403.12345] Paper Title Here\r
\r
**Authors:** John Doe, Jane Smith\r
**Category:** cs.AI\r
**Submitted:** 2024-03-15\r
\r
**Abstract:**\r
This paper presents a novel approach to...\r
\r
**Links:**\r
- arXiv: https://arxiv.org/abs/2403.12345\r
- PDF: https://arxiv.org/pdf/2403.12345.pdf\r
\r
---\r
```\r
\r
Starred papers include a ⭐ marker for easy identification.\r
\r
## Workflow\r
\r
When user asks to check arXiv:\r
\r
1. **Identify categories** - Parse user request for category names\r
2. **Fetch papers** - Run `arxiv_watcher.py fetch \x3Ccategory>`\r
3. **Format output** - Present papers in readable Markdown\r
4. **Offer actions** - Ask if user wants to star any papers\r
\r
### Example Interactions\r
\r
**User:** "check new papers in cs.AI"\r
\r
**Assistant:** Runs fetch command, displays 5-10 recent papers, asks if any should be starred.\r
\r
**User:** "star the second one"\r
\r
**Assistant:** Parses the arxiv_id from previous output, runs star command, confirms action.\r
\r
**User:** "show my starred papers"\r
\r
**Assistant:** Runs starred command, displays list with details.\r
\r
## Resources\r
\r
### scripts/\r
- `arxiv_watcher.py` - Main CLI for fetching and managing papers\r
\r
### assets/\r
- `starred.json` - Persistent storage for starred papers (auto-created)\r
安全使用建议
This skill appears coherent and does what it claims: it fetches arXiv Atom feeds (export.arxiv.org / arxiv.org), formats results, and stores starred papers in assets/starred.json. It requires no credentials and has no install steps. Before installing or running: 1) confirm you trust the skill source (no homepage provided); 2) note it needs outbound network access to arxiv.org and will create/write assets/starred.json in the skill directory (avoid running in a sensitive/shared directory if you have concerns); 3) ensure a suitable Python environment is available; and 4) review the full script file locally (the provided file listing here was truncated at the end, so verify the script's complete main() dispatch and any remaining lines before granting runtime access).
功能分析
Type: OpenClaw Skill
Name: arxiv-watch
Version: 1.0.0
The arxiv-watch skill is a legitimate tool designed to fetch and track research papers from arXiv.org. The Python script (scripts/arxiv_watcher.py) uses standard libraries to query the official arXiv API and manages a local bookmarking system via a JSON file (assets/starred.json). There is no evidence of data exfiltration, malicious execution, or prompt injection attempts.
能力评估
Purpose & Capability
Name/description, SKILL.md, and the included script all focus on fetching arXiv feeds, formatting results, and managing a local starred list. The requested capabilities (none) match what the skill does. No unrelated credentials, binaries, or config paths are required.
Instruction Scope
SKILL.md instructs the agent to run the bundled CLI to fetch papers and manage starred entries. The runtime behavior described and the script itself only read/write the local assets/starred.json and perform HTTP requests to arxiv.org; there are no instructions to read unrelated files or exfiltrate data to third-party endpoints.
Install Mechanism
There is no install spec (instruction-only). A Python script is included and relies on stdlib modules; nothing is downloaded from external or unknown URLs and no archives are extracted.
Credentials
The skill requests no environment variables or secrets. It performs plain HTTP(S) requests to arxiv.org only and stores starred metadata locally in assets/starred.json. No unrelated credentials or high-privilege paths are requested.
Persistence & Privilege
always is false and the skill does not request persistent platform-wide privileges. It writes only its own assets/starred.json and does not modify other skills or system-wide configurations.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install arxiv-watch - 安装完成后，直接呼叫该 Skill 的名称或使用
/arxiv-watch触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
init
元数据
常见问题
Arxiv Watcher 是什么？
Fetch and monitor latest arXiv papers by category, generate summaries, and manage starred papers for easy reference. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 281 次。
如何安装 Arxiv Watcher？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install arxiv-watch」即可一键安装，无需额外配置。
Arxiv Watcher 是免费的吗？
是的，Arxiv Watcher 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Arxiv Watcher 支持哪些平台？
Arxiv Watcher 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Arxiv Watcher？
由 constx1337（@constx1337）开发并维护，当前版本 v1.0.0。
推荐 Skills