dsh-scholar DSH plugin

This plugin adds three tools to DSH: arxiv_search for searching arXiv papers, scholar_fetch for fetching full text from URLs including PDF extraction, and journal_lookup for querying journal metadata via OpenAlex. It addresses the lack of PDF support and proxy configuration in DSH's native web_fetch tool.

skills-workflows developer-tools

How to install the dsh-scholar DSH plugin

dsh plugin add dsh-scholar

Copying does not run this command. Review the repository and version before installing the dsh-scholar DSH plugin.

dsh-scholar DSH plugin data source

dsh-scholar DSH plugin snapshot date: Aug 16, 2026

discovered

What the dsh-scholar DSH plugin can do

Search arXiv papers by title, author, abstract, returning links to abs and PDF.
Fetch full text from any HTTP(S) URL, converting HTML to plain text and extracting text from PDF.
Query journal metadata including publisher, ISSN, total papers, citations, and citation indicators using OpenAlex API.
Configure HTTP proxy for accessing foreign paper sites (e.g., arXiv).
Adjustable settings: timeout, max results for arXiv, and individual tool enable/disable.

Where the dsh-scholar DSH plugin fits

Rapidly search for relevant papers on arXiv during literature review.
Download and extract full-text content of papers for automated analysis by an agent.
Look up journal details to decide where to submit a manuscript.
Integrate with DSH workflows to build a research assistance pipeline.

Who the dsh-scholar DSH plugin is for

Researchers and academics who need to search and fetch papers programmatically.
Developers building AI agents for scientific literature analysis.
Students and scholars using DSH as a personal research assistant.

dsh-scholar DSH plugin limitations

Journal_lookup metrics from OpenAlex are not official JCR Impact Factors; use official sources for critical decisions.
PDF text extraction quality depends on the PDF itself; scanned documents or formula-heavy layouts may produce poor results, especially for mathematical formulas.
Dependent on third-party APIs (arXiv and OpenAlex); users must comply with their terms and rate limits.
Proxy configuration is required for users in China or behind restrictive networks to access some paper sites.

dsh-scholar DSH plugin: from the repository README

Quoted from the 202411130091-coder/dsh-scholar README, the upstream source of the dsh-scholar DSH plugin. Copyright remains with the original authors.

DeepSeek Harness（DSH）科研工具套件：给 agent 补上「检索论文 / 读论文全文 / 查期刊」三个工具。 ## 提供的工具 | 工具 | 作用 | 数据源 | |---|---|---| | `arxiv_search` | 搜 arXiv 论文（标题/作者/摘要/abs+pdf 链接） | 官方免费 arXiv API，无需 key | | `scholar_fetch` | 按 URL 抓全文：HTML→纯文本、**PDF→抽文本** | 任意 HTTP(S)，可选代理 | | `journal_lookup` | 查期刊元数据（出版社/ISSN/发文量/被引/引用指标） | OpenAlex API，无需 key | ## 为什么做这个 DSH 原生的 `web_fetch` 只能抓 HTML/纯文本，**不支持 PDF**、也**不带代理**；对科研场景（读 arXiv PDF、查期刊）是明显缺口。`dsh-scholar` 补上这两点。 ## 安装 ```bash dsh plugin add dsh-scholar ``` 装完在组合里自动插入一个 host 行，三个工具默认全开。需要代理（国内访问境外论文站）时，编辑 profile 的 `cordis.patch.yml`，给 `scholar` 行加： ```yaml - id: scholar config: proxy: http://127.0.0.1:7890 # 你的 Clash 混合端口；留空 = 直连 ``` ## 配置项 | 字段 | 默认 | 说明 | |---|---|---| | `arxiv` / `fetch` / `journal` | `true` | 三个工具的开关 | | `proxy` | `""` | HTTP(S) 出网代理 URL，空 = 直连 | | `fetchTimeoutMs` | `30000` | 单次请求超时预算 | | `arxivMaxResults` | `8` | arXiv 单次最多返回条数 | ## 注意 - **`journal_lookup` 返回的 OpenAlex 引用指标不是官方 JCR 影响因子**。判断分区/影响因子请以官方（JCR / 中科院分区）为准。 - 本项目采用 MIT 许可证；对第三方数据源（arXiv / OpenAlex）的调用请遵守其使用条款与频率限制。 - PDF 抽取质量取决于 PDF 本身（扫描版/公式密集的排版可能抽得差，公式会碎）；数学公式以 LaTeX 源为最佳。 ## 开发 ```bash npm install # 装 unpdf / undici npm link # 本地联调 ```

Read the full READMERepository license: MIT

dsh-scholar DSH plugin questions

How do I install dsh-scholar?

Run `dsh plugin add dsh-scholar` in your terminal. After installation, a host line will be automatically inserted into your profile, enabling all three tools by default.

How can I configure a proxy for fetching papers?

Edit your profile's `cordis.patch.yml` file and add a `proxy` field under the `scholar` host config, for example `proxy: http://127.0.0.1:7890`. Leave it empty to use direct connection.

Does dsh-scholar require any API keys?

No, both arXiv search and journal lookup use free public APIs (arXiv API and OpenAlex API) that do not require an API key. The full-text fetcher can access any HTTP(S) URL without authentication.

Can I use dsh-scholar to fetch papers from sources other than arXiv?

The `scholar_fetch` tool can fetch full text from any URL, not just arXiv. However, the search tool only searches arXiv. For non-ArXiv papers, you need to provide a URL manually.

How accurate is the PDF text extraction?

The extraction quality depends on the PDF. It works well for text-based PDFs but may produce poor results for scanned documents or formula-heavy content. Mathematical formulas may be fragmented. For best results, use LaTeX source when available.