cnki-paper-detail

name	cnki-paper-detail
description	Extract full paper details from a CNKI paper page including title, authors, affiliations, abstract, keywords, fund, classification. Use when the user needs detailed information about a specific paper.
argument-hint	[paper URL or blank if already on detail page]

CNKI Paper Detail Extraction

RC browser tool: navigate with browser action=open url="..."; run the JS in each step via browser action=act kind=evaluate fn="<the async function shown>" (the whole function body goes into fn). Never pass a profile — RC uses its default managed Chrome (CDP 18800).

Extract complete metadata from a CNKI paper detail page.

Arguments

$ARGUMENTS is optionally a CNKI paper detail URL (containing kcms2/article/abstract). If not provided, assumes the current page is already a paper detail page.

Steps

1. Navigate to the paper page (if URL provided)

If $ARGUMENTS contains a URL:

• Use browser action=open with the URL.
• Use browser action=act kind=wait with text ["摘要"] and timeout 15000.

2. Check for captcha

Use browser action=snapshot mode=efficient. If "拖动下方拼图完成验证" found, notify user:

CNKI 正在显示滑块验证码。请在 Chrome 浏览器中手动完成拼图验证，完成后告诉我继续。

3. Extract paper metadata via JavaScript

Use browser action=act kind=evaluate with this function:

() => {
  const brief = document.querySelector('.brief');
  if (!brief) return { error: 'Paper detail section (.brief) not found' };

  // Title
  const title = brief.querySelector('h1')?.innerText?.trim()
    ?.replace(/\s*附视频\s*$/, '')  // remove "附视频" suffix
    ?.replace(/\s*网络首发\s*$/, ''); // remove "网络首发" suffix

  // Authors - first h3.author contains author links with sup tags
  const authorH3s = brief.querySelectorAll('h3.author');
  const authorSection = authorH3s[0];
  const authors = [];
  if (authorSection) {
    const authorLinks = authorSection.querySelectorAll('a');
    authorLinks.forEach(a => {
      // Affiliation superscripts live in a <sup> (e.g. "曲文<sup>1,2</sup>").
      // Read the number(s) from <sup> and the name from the first text node so
      // multi-affiliation authors ("1,2") and trailing icons (<i class="icon-email">) don't corrupt the name.
      const supEl = a.querySelector('sup');
      const affiliationNum = supEl ? supEl.innerText.() : ;
       textNode = .(a.).( n. ===  && n..());
       name = textNode ? textNode..() : a..();
       (!textNode && affiliationNum && name.(affiliationNum)) {
        name = name.(, name. - affiliationNum.).();
      }
       (name) authors.({ name, affiliationNum });
    });
  }

  
   affiliations = [];
   (authorH3s. > ) {
     orgLinks = authorH3s[].();
    orgLinks.( {
      affiliations.(a.?.());
    });
  }

  
   abstractEl = .();
   abstract = abstractEl?.?.() || ;

  
   keywordsP = .();
   keywords = keywordsP
    ? .(keywordsP.()).( a.?.(, ).())
    : [];

  
   fundsP = .();
   fund = fundsP?.?.() || ;

  
   clcCode = .();
   classification = clcCode?.?.() || ;

  
   docTop = .();
   journal = docTop?.()?.?.() || ;

  
   headTime = .();
   pubInfo = headTime?.?.() || ;

  
   isOnlineFirst = !!brief.();

  
   catalogList = .();
   toc = catalogList?.?.() || ;

  
   citationTabs = .();
   citationInfo = {};
  citationTabs.( {
     id = li.();
     text = li.?.();
     countMatch = text.();
     (id) {
      citationInfo[id] = {
        : text.(, ).(),
        : countMatch ? (countMatch[]) : 
      };
    }
  });

   {
    title,
    authors,
    affiliations,
    abstract,
    keywords,
    fund,
    classification,
    journal,
    pubInfo,
    isOnlineFirst,
    toc,
    citationInfo
  };
}

4. Format and present the output

## {title} {isOnlineFirst ? "[网络首发]" : ""}

**Authors:**
{For each author: "- {name} ({affiliation})"}

**Affiliations:**
{For each affiliation: "- {affiliation}"}

**Journal:** {journal}
**Publication Info:** {pubInfo}

**Abstract:**
{abstract}

**Keywords:** {keywords joined by ", "}

**Fund:** {fund}
**Classification:** {classification}

**Citation Network:**
{For each citation type: "- {label}: {count}"}

5. Fallback: snapshot-based parsing

If JS extraction fails, use browser action=snapshot mode=efficient and parse the accessibility tree:

• Title: heading level 1 element
• Authors: link elements whose URLs contain kcms2/author/detail
• Affiliations: link elements whose URLs contain kcms2/organ/detail
• Abstract: StaticText following "摘要："
• Keywords: link elements whose URLs contain kcms2/keyword/detail
• Fund: link elements following "基金资助："
• Classification: StaticText following "分类号："

Verified DOM Selectors

Data	Selector	Notes
Paper section	.brief	Main paper info container
Title	.brief h1	May contain icons, clean text needed
Authors	.brief h3.author:first-of-type a	Name in first text node; affiliation #s in child <sup> (e.g., 张三<sup>1,2</sup>)
Affiliations	.brief h3.author:nth-of-type(2) a	Text starts with "N." (e.g., "1.北京大学")
Abstract	.abstract-text	Full abstract text
Keywords	p.keywords a	Semicolon-separated keyword links
Fund	p.funds	Fund information text
Classification	.clc-code	CLC classification codes
Journal	.doc-top a	Source journal link
Online first	.brief .icon-shoufa	Present if paper is online first
Citation tabs	ul.module-tab.tpl_lieteratures li	data-id attr identifies type