Agent skill

Nature Academic Search

by wp-a in wp-a/nature-academic-search

A skill your agent uses when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID, 或导出 RIS、BibTeX、NBIB、ENW;also use for…

MITAuto-check passedResearch & Science

Install Nature Academic Search

skills CLI
$ npx skills add wp-a/nature-academic-search --skill nature-academic-search -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install wp-a/nature-academic-search nature-academic-search --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
nature-academic-search
GitHub stars
304
Token cost
~1.4k tokens
SKILL.md length
329 words
Files
171 (incl. scripts, references)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

A skill your agent uses when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID, 或导出 RIS、BibTeX、NBIB、ENW;also use for…

  • Works in 4 steps: 中文问题先 lookup_mesh… → search_papers(建议 ranking="relevance");必须报告 → 对拟引用记录(通常先 3 条)调用 get_paper_by_id +… → …
  • Citation verification
  • SKILL.md covers 什么时候触发, 任务路由, 最小成功路径 and 来源与边界, plus 5 more sections
  • Reaches api.wpironman.top; needs ACADEMIC_SEARCH_LLM_API_KEY

What it does

Nature Academic Search is an agent skill from wp-a/nature-academic-search. Use when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID, 或导出 RIS、BibTeX、NBIB、ENW;also use for multi-source academic search, citation verification, citation graphs, trial registration search, research workflows, or DeepSeek Harness academic-search MCP setup.

Its SKILL.md is about 1.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 177 other files, including scripts and reference files (for example `.agents/plugins/marketplace.json`, `.claude-plugin/marketplace.json` and `.github/dependabot.yml`).

It sits in Research & Science, covering Academic paper search. It works with arXiv, PubMed, Model Context Protocol and DeepSeek. The repository describes itself as: Academic Paper Search:中文科研用户的 Codex / Claude Code / DeepSeek Harness Skill + MCP;跨 CrossRef、PubMed、arXiv、OpenAlex、Europe PMC 检索去重,支持 MeSH、引用核验、引文图谱、试验检索与审计导出。 The licence is MIT.

When your agent uses it

  • Citation verification
  • Citation graphs
  • Trial registration search
  • Research workflows

Example prompts

  • “/nature-academic-search”

Requirements

  • A credential in ACADEMIC_SEARCH_LLM_API_KEY

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

  1. 中文问题先 lookup_mesh 或写出英文检索式;search_run.query_analysis.mesh_required
  2. search_papers(建议 ranking="relevance");必须报告
  3. 对拟引用记录(通常先 3 条)调用 get_paper_by_id + expected。
  4. 按 verified / mismatch / not_found / manual_needed 分组交付。

What it can do on your machine

Read from SKILL.md and the folder at commit 725c2ec. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/, which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Hosts in commands or code, which the agent is likely to contact:

    • api.wpironman.top

    Also links to:

    • github.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • ACADEMIC_SEARCH_LLM_API_KEY

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Nature Academic Search loads about 1.4k tokens when it runs, and up to ~14k if it reads all its reference files. Until then it costs about 90 tokens; SKILL.md has 329 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~90
When it runs · the whole SKILL.md, loaded when a task matches
~1.4k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~14k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from wp-a/nature-academic-search at commit 725c2ec, republished under its MIT licence (© wp-a). 329 words, ~1,407 tokens.

Download SKILL.mdSave it as .claude/skills/nature-academic-search/SKILL.md (or your agent's skills folder). This skill also uses 170 other files; get the full folder from GitHub.
name
nature-academic-search
description
Use when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID, 或导出 RIS、BibTeX、NBIB、ENW;also use for multi-source academic search, citation verification, citation graphs, trial registration search, research workflows, or DeepSeek Harness academic-search MCP setup.

什么时候触发

用户提到找文献、文献检索、引用核验、引文图谱、MeSH、PubMed、预印本、临床试验、RIS/BibTeX 或 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT 时触发本 skill。目标是论文时使用 entity_type="publication";目标是注册试验时使用 entity_type="trial",两者不能按题名合并。

任务路由

以下工具路由以实际工具输出为准;适配器存在不代表本次已经查询。

用户目标入口执行要求
找文献、综述前期检索search_papers按源检索、去重,报告 sources_queried / sources_succeeded / sources_skipped / errors
核验已有引用get_paper_by_id + expected输出字段级 verified、mismatch、not_found 或 manual_needed
追踪上下游引文get_paper_by_id(include_relations=true)默认 depth=1;需要二跳时显式 depth=2,保留图谱边和来源缺口
构建 PubMed 检索式lookup_mesh先核对 MeSH 词和 ID,再组合自由词;不要凭空猜主题词
生成单条引用get_citation只格式化已解析论文;NCT 注册不生成论文引用
批量导出或自动化CLI citation / workflow保留 run.json、核验状态和人工待处理清单

客户端可能给工具名添加 MCP 前缀;工具总数仍为四个,不新增专用图谱工具。 DeepSeek Harness 使用独立的 dsh-academic-paper-search Bundle,通过官方 @deepseek-ai/dsh-mcp-client 映射为 mcp__academic_search__*;这是同一 MCP 运行时的客户端适配,不要在 DSH 中重写或假设额外的论文工具。 CLI 从 0.3.1 起可用 nature-academic-search search / verify 得到同一份 JSON; 旧版客户端先检查 --help,不要假设已安装包包含新命令。

最小成功路径

  1. 中文问题先 lookup_mesh 或写出英文检索式;search_run.query_analysis.mesh_required 为 true 时不得跳过。不要把中文整句直接当已核验检索。
  2. search_papers(建议 ranking="relevance");必须报告 sources_queried / sources_succeeded / sources_skipped / errors。
  3. 对拟引用记录(通常先 3 条)调用 get_paper_by_id + expected。
  4. 按 verified / mismatch / not_found / manual_needed 分组交付。

正确:lookup_mesh("medical education") → search_papers("generative AI medical education", ranking="relevance") → 对 DOI 做 expected 核验。 错误:只调用 search_papers("生成式AI医学教育"),并把返回题名当作已核验参考文献。

来源与边界

  • 默认论文源:crossref、pubmed、arxiv、openalex、europe_pmc。
  • semantic_scholar 用于显式搜索、enrich 或图谱;富化优先使用 DOI、PMID、arXiv 等强标识符。
  • OpenAlex 提供跨学科 references / cited_by;PubMed ELink 提供生物医学双向关系。
  • Crossref 与 Europe PMC 主要提供 references;arXiv 目前只作为节点和版本线索。
  • clinicaltrials_gov 只服务 entity_type="trial",不是论文数据库。
  • 未连接 Google Scholar、Web of Science、Scopus、Embase、CNKI、万方,不得声称覆盖这些来源。

源失败、限流、缺少强标识符或不支持某方向时,保留成功结果并记录缺口;不能把未查询到解释成没有关系。

引文图谱契约

调用 get_paper_by_id 时可以传:

json
{
  "include_relations": true,
  "relation": "both",
  "depth": 1,
  "rows": 20,
  "relation_sources": ["openalex", "crossref", "pubmed", "europe_pmc", "semantic_scholar"]
}

relation 是 references、cited_by 或 both。输出 citation_graph 必须保留:

  • nodes:合并后的 publication 节点和来源记录;
  • edges:统一为 citing → cited,含 relation 与 observed_by;
  • sources_queried、sources_succeeded、sources_skipped、errors;
  • truncated、truncation_reason、depth_completed。

references 表示种子指向被引用节点,cited_by 表示引用者指向种子。图谱是关系导航和审计数据, 不是证据质量、因果关系、影响力或研究结论评分。

标准执行顺序

  1. 明确主题、人群/系统、干预、结局、日期、文献类型、预印本政策和实体类型。
  2. 生物医学问题先用 lookup_mesh 核验主题词,再组合题名/摘要自由词。含中文的问题必须先得到 MeSH 或英文检索式。
  3. 调用 search_papers;保存原查询、日期、请求源、结果数量和 search_run。
  4. 按 DOI、PMID、PMCID、arXiv、OpenAlex、Semantic Scholar 或 NCT 强标识符去重;弱题名匹配保留冲突。
  5. 对拟引用记录调用 get_paper_by_id + expected,逐项核对题名、首位作者、年份、期刊和标识符。
  6. 需要时追加一跳引文图谱;二跳、rows 和源列表必须有明确研究目的和预算。
  7. 按 verified、mismatch、not_found、manual_needed、preprint、trial 分组交付;未核验记录不能静默导出。
  8. 批量任务使用 workflow,先生成 plan.json,获得批准后再检索,并保存 run.json、results.json、 verification.json、screening.csv、references.ris 和可选 graph.json。

Workflow 从 0.3.1 起默认按标识符回查并核验候选元数据。省略 verify 不会把候选视为已核验。 核验范围见 fields、lookup_id / lookup_id_type;来源无法提供的附加 ID 列入 unchecked_identifiers,不能称为已核验,RIS 不写入未匹配的 DOI / PMID。作者截断标记 et al. 不参与作者比较。原始候选和未核验附加 ID 保留在 results.json。 导出默认要求 verified;有 screen 时还要求筛选为 include,exclude、pending_manual、 缺失/无效决定和模型失败的待处理记录均不导出。无需模型筛选时使用 steps: [plan, search, verify, export]。保留人工待处理记录,报告 exported_count; trial 只保留注册数据,不写成论文 RIS。citation CLI 只下载/转换格式,不自动做 expected 核验。

WPIRONMAN 中转

WPIRONMAN 是可选的 OpenAI-compatible 模型入口,当前 runner 用于摘要级 screen; plan.json 根据 YAML 本地生成,研究计划或规则可先在客户端整理。 它不是论文来源、数据库、引用验证器,也不替代 Crossref、PubMed、OpenAlex、Europe PMC 或 Semantic Scholar。

bash
export ACADEMIC_SEARCH_LLM_BASE_URL=https://api.wpironman.top/v1
export ACADEMIC_SEARCH_LLM_API_KEY=你的中转密钥
export ACADEMIC_SEARCH_LLM_MODEL=你的模型名
export ACADEMIC_SEARCH_LLM_PROTOCOL=responses_http

默认只发送标题、摘要、标识符和获准元数据;全文必须显式设置 privacy.allow_full_text: true。 密钥不得进入日志、manifest 或 prompt artifact。中转超时、限流或返回坏 JSON 时, 失败的模型步骤标记为 skipped,学术检索、核验和审计继续;待人工筛选记录不进入 RIS。

结果契约

每次成功的 search_papers 返回 search_run,至少保存 run_id、UTC 时间、请求参数、去重前后数量和 result_fingerprint。每条记录保留稳定 record_id、sources、source_records、conflicts、 citation_counts 和 citation_count_source。不得合成没有来源的“总引用数”。

统一 filters 支持日期、语言、作者、文献类型和强标识符;ranking 可设为 relevance 或 none。 相关性排序会写入 ranking_score、ranking_reasons 和 score_version,只表示检索相关性,不表示证据质量。 含中文的查询会在 query_analysis 中给出 contains_cjk、latin_terms、cjk_terms 和 mesh_required。

详细查询构建、来源分层、引用文件和工作流见:

证据规则

  • 不编造元数据、摘要、引用次数、标识符、开放获取状态、全文结论或试验结果。
  • 不把预印本描述为同行评审论文,不把 trial 注册描述为已发表研究。
  • 不因为来源适配器存在就声称它本次已查询;以实际 sources_queried 为准。
  • 不把引用数量、图谱度数或模型排序当作证据质量或因果证据。

© wp-a, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 170 other files (scripts, references) in the repository root of wp-a/nature-academic-search.

  • SKILL.md
  • .agents/plugins/marketplace.json
  • .claude-plugin/marketplace.json
  • .github/dependabot.yml
  • .github/workflows/ci.yml
  • .github/workflows/network-smoke.yml
  • .github/workflows/publish.yml
  • .gitignore
  • LICENSE
  • README.md
  • config/mcp-snippet.json
  • config/settings-snippet.json
  • config/triggers-academic-search.toml
  • docs/assets
  • … and 157 more

Open the folder on GitHubat commit 725c2ec

Compare with similar skills

Nature Academic Search next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Nature Academic Search compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Nature Academic Search this skillwp-a/nature-academic-search304—~1.4kAutomated safety check: PassMIT
Nature Academic Searchjing1312/nature-figure-skill171—~1.3kAutomated safety check: NotesMIT
Ai4scholar ResearchDrchronx/ai-agent-research-starter-kit139—~1.4kAutomated safety check: PassCustom licence
Paper Searchopenags/paper-search-mcp2.8k—~1.2kAutomated safety check: NotesMIT
Paper Searchopenags/paper-search-mcp2.8k—~794Automated safety check: NotesMIT
Paper Searchdr-dumpling/paper-search-cli135—~1.1kAutomated safety check: PassMIT

Similar skills

  • Nature Academic Search

    jing1312/nature-figure-skill

    Multi-source literature search, citation verification, MeSH search strategy, citation file management (.nbib/.ris/.bib conversion), and reference management (BibTeX, related articles, ID conversion)…

    171 GitHub stars~1.3k tokensUpdated 1 mo ago
    Research & ScienceAuto-check: notes
  • Ai4scholar Research

    Drchronx/ai-agent-research-starter-kit

    Use AI4Scholar for real scholarly literature tasks through MCP or the OpenClaw plugin: paper search, Google Scholar/Semantic Scholar/PubMed/arXiv/bioRxiv/medRxiv queries, PDF download and reading…

    139 GitHub stars~1.4k tokensUpdated 4 mo ago
    Research & ScienceAuto-check passed
  • Paper Search

    openags/paper-search-mcp

    Search, download, and read academic papers from 20+ sources (arXiv, PubMed, Semantic Scholar, CrossRef, etc).

    2.8k GitHub stars~1.2k tokensUpdated 9 days ago
    Research & ScienceAuto-check: notes
  • Paper Search

    openags/paper-search-mcp

    Search, download, and read academic papers through the paper-search MCP tools (mcppaper-search), covering arXiv, PubMed, bioRxiv, Semantic Scholar, Crossref, OpenAlex and 15+ other sources.

    2.8k GitHub stars~794 tokensUpdated 9 days ago
    Research & ScienceAuto-check: notes
  • Paper Search

    dr-dumpling/paper-search-cli

    学术文献检索与论文获取调度器,基于 paper-search CLI,而不是 MCP server. An agent skill from dr-dumpling/paper-search-cli.

    135 GitHub stars~1.1k tokensUpdated 3 mo ago
    Research & ScienceAuto-check passed
  • Literature Review

    Norman-bury/research-writing-skill

    A skill your agent uses when writing literature review sections - guides searching, organizing, and synthesizing academic sources

    3.4k GitHub stars~2.2k tokensUpdated 4 mo ago
    Research & ScienceAuto-check: notes

Questions about Nature Academic Search

What does Nature Academic Search do?

A skill your agent uses when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID, 或导出 RIS、BibTeX、NBIB、ENW;also use for…. Nature Academic Search is an agent skill from wp-a/nature-academic-search. Use when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID, 或导出 RIS、BibTeX、NBIB、ENW;also use for multi-source academic search, citation verification, citation graphs, trial registration search, research workflows, or DeepSeek Harness academic-search MCP setup.

When should I use Nature Academic Search?

Nature Academic Search fits situations like: citation verification; citation graphs; trial registration search; research workflows.

How do I install Nature Academic Search in Claude Code?

Run `npx skills add wp-a/nature-academic-search --skill nature-academic-search -a claude-code`. Or copy the skill folder (the wp-a/nature-academic-search repository) into .claude/skills/nature-academic-search in your project. Claude Code loads it when a task matches its description.

How do I install Nature Academic Search in Codex?

Run `npx skills add wp-a/nature-academic-search --skill nature-academic-search -a codex`. Or copy the skill folder (the wp-a/nature-academic-search repository) into .agents/skills/nature-academic-search in your project. Codex loads it when a task matches its description.

Can I use Nature Academic Search in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add wp-a/nature-academic-search --skill nature-academic-search -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/nature-academic-search, .gemini/skills/nature-academic-search, .github/skills/nature-academic-search and .opencode/skills/nature-academic-search in your project.

What does Nature Academic Search need to run?

Going by SKILL.md and its folder, Nature Academic Search needs credentials named ACADEMIC_SEARCH_LLM_API_KEY. Our summary lists: A credential in ACADEMIC_SEARCH_LLM_API_KEY.

Does Nature Academic Search access the network?

SKILL.md names 2 domains. In commands or code: api.wpironman.top; the agent is likely to contact it when it follows the instructions. As links in the text: github.com. This is read from the text; nothing was executed.

Is Nature Academic Search safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Nature Academic Search use?

Nature Academic Search is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Nature Academic Search use?

About 1.4k tokens (SKILL.md is roughly 5.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 12k tokens, read only when the agent opens those files.

What are the alternatives to Nature Academic Search?

Skills that share tags, products or a category with Nature Academic Search: Nature Academic Search (jing1312/nature-figure-skill, 171 stars), Ai4scholar Research (Drchronx/ai-agent-research-starter-kit, 139 stars), Paper Search (openags/paper-search-mcp, 2.8k stars) and Paper Search (openags/paper-search-mcp, 2.8k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Nature Academic Search?

wp-a (a GitHub user) maintains it in wp-a/nature-academic-search, which has 304 GitHub stars. The repository was last updated on October 1, 2026.

Source: wp-a/nature-academic-search on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.