mcp%3Asetup-arxiv-mcp by neolabhq/context-engineering-kit
npx skills add https://github.com/neolabhq/context-engineering-kit --skill mcp:setup-arxiv-mcp$ARGUMENTS
询问用户希望将配置存储在何处:
选项:
./CLAUDE.md./CLAUDE.local.md
* 请确认这些文件已列在 .gitignore 中,若未列出则添加~/.claude/CLAUDE.md存储用户的选择,并在后续步骤中使用相应的路径。
首先,尝试使用 mcp-find 工具搜索服务器,以验证 Docker MCP (MCP_DOCKER) 是否可访问。
如果 Docker MCP 不可用:
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
告知用户,已知常规的 arxiv-mcp-server 存在问题,特别是初始化失败(初始化期间出现 EOF 错误)。因此,我们将改用 paper-search MCP 服务器。
使用 Docker MCP 查找并添加提供全面学术论文搜索功能的 paper-search MCP 服务器:
mcp-find query: "paper-search"
mcp-add name: "paper-search" activate: true
此服务器提供对多个学术来源的访问:
通过搜索论文来验证服务器是否正常工作:
mcp-exec name: "search_arxiv" arguments: {"query": "test query", "max_results": 2}
使用步骤 1 中确定的路径:
成功设置 paper-search MCP 服务器后,使用以下内容更新 CLAUDE.md 文件:
### 使用 Paper Search MCP 进行学术研究
可通过 Docker MCP 使用 Paper Search MCP 来搜索和下载学术论文。
**可用工具**:
- `search_arxiv` - 搜索 arXiv 预印本(物理、数学、计算机科学等)
- `search_pubmed` - 搜索 PubMed 生物医学文献
- `search_biorxiv` / `search_medrxiv` - 搜索生物学/医学预印本
- `search_semantic` - 搜索 Semantic Scholar(支持年份筛选)
- `search_google_scholar` - 广泛的学术搜索
- `search_iacr` - 搜索密码学论文
- `search_crossref` - 通过 DOI/引文搜索
**下载和阅读工具**:
- `download_arxiv` / `read_arxiv_paper` - 下载/阅读 arXiv PDF
- `download_biorxiv` / `read_biorxiv_paper` - 下载/阅读 bioRxiv PDF
- `download_semantic` / `read_semantic_paper` - 通过 Semantic Scholar 下载/阅读
**使用说明**:
- 使用 `mcp-exec` 调用工具,例如:`mcp-exec name: "search_arxiv" arguments: {"query": "topic", "max_results": 10}`
- 默认情况下,下载的论文会保存到 `./downloads`
- 对于 Semantic Scholar,支持多种 ID 格式:DOI、ARXIV、PMID 等。
如果您特别需要具有额外功能(深度分析提示、本地存储管理)的专用 arXiv MCP 服务器,可以尝试:
mcp-find query: "arxiv"
mcp-config-set server: "arxiv-mcp-server" key: "storage_path" value: "/path/to/papers"
mcp-add name: "arxiv-mcp-server" activate: true
注意:此服务器需要为下载的论文配置存储路径。
每周安装次数
152
代码仓库
GitHub 星标数
699
首次出现
2026年2月19日
安装于
opencode146
github-copilot145
codex145
gemini-cli144
kimi-cli142
cursor142
User Input:
$ARGUMENTS
Ask the user where they want to store the configuration:
Options:
Project level (shared via git) - Configuration tracked in version control, shared with team
./CLAUDE.mdProject level (personal preferences) - Configuration stays local, not tracked in git
./CLAUDE.local.md.gitignore, add them if notUser level (global) - Configuration applies to all projects for this user
~/.claude/CLAUDE.mdStore the user's choice and use the appropriate paths in subsequent steps.
First, verify that Docker MCP (MCP_DOCKER) is accessible by attempting to use mcp-find tool to search for servers.
If Docker MCP is NOT available:
Write to user that regular arxiv-mcp-server is known to have issues, specifically is failing to initialize (EOF error during init). So we will use paper-search MCP server instead.
Use Docker MCP to find and add the paper-search MCP server which provides comprehensive academic paper search capabilities:
mcp-find query: "paper-search"
mcp-add name: "paper-search" activate: true
This server provides access to multiple academic sources:
Verify the server is working by searching for papers:
mcp-exec name: "search_arxiv" arguments: {"query": "test query", "max_results": 2}
Use the path determined in step 1:
Once the paper-search MCP server is successfully set up, update CLAUDE.md file with the following content:
### Use Paper Search MCP for Academic Research
Paper Search MCP is available via Docker MCP for searching and downloading academic papers.
**Available tools**:
- `search_arxiv` - Search arXiv preprints (physics, math, CS, etc.)
- `search_pubmed` - Search PubMed biomedical literature
- `search_biorxiv` / `search_medrxiv` - Search biology/medicine preprints
- `search_semantic` - Search Semantic Scholar with year filters
- `search_google_scholar` - Broad academic search
- `search_iacr` - Search cryptography papers
- `search_crossref` - Search by DOI/citation
**Download and read tools**:
- `download_arxiv` / `read_arxiv_paper` - Download/read arXiv PDFs
- `download_biorxiv` / `read_biorxiv_paper` - Download/read bioRxiv PDFs
- `download_semantic` / `read_semantic_paper` - Download/read via Semantic Scholar
**Usage notes**:
- Use `mcp-exec` to call tools, e.g., `mcp-exec name: "search_arxiv" arguments: {"query": "topic", "max_results": 10}`
- Downloaded papers are saved to `./downloads` by default
- For Semantic Scholar, supports multiple ID formats: DOI, ARXIV, PMID, etc.
If you specifically need the dedicated arXiv MCP server with additional features (deep analysis prompts, local storage management), you can try:
mcp-find query: "arxiv"
mcp-config-set server: "arxiv-mcp-server" key: "storage_path" value: "/path/to/papers"
mcp-add name: "arxiv-mcp-server" activate: true
Note: This server requires configuration of a storage path for downloaded papers.
Weekly Installs
152
Repository
GitHub Stars
699
First Seen
Feb 19, 2026
Installed on
opencode146
github-copilot145
codex145
gemini-cli144
kimi-cli142
cursor142
Azure 升级评估与自动化工具 - 轻松迁移 Functions 计划、托管层级和 SKU
94,100 周安装