This project provides a set of Claude Code CLI skills to programmatically interact with CNKI (中国知网), China's largest academic database. It enables users to search for papers, browse journals, download PDFs, and export citations to Zotero, streamlining academic research workflows directly from the command line. The target audience includes researchers and power users leveraging Claude Code for literature review and data acquisition.

How It Works

The core approach utilizes Chrome DevTools Protocol (MCP) to execute JavaScript directly within a running Chrome browser instance, bypassing the need for brittle screen scraping or OCR. Each skill is designed as a single asynchronous evaluate_script call, ensuring fast and reliable interactions. Key design choices include direct page navigation over link clicking to manage browser tabs efficiently, batch export capabilities for multiple papers, and built-in detection for captchas, which pauses execution for manual user resolution.

Quick Start & Requirements

1. Install Chrome DevTools MCP: claude mcp add chrome-devtools -- npx -y chrome-devtools-mcp@latest
2. Clone the repository and copy skills/agents to your Claude configuration: git clone https://github.com/cookjohn/cnki-skills.git followed by cp -r skills/ agents/ .claude/.
3. Launch Chrome with remote debugging enabled: chrome --remote-debugging-port=9222.
4. Start Claude Code CLI; skills are automatically detected.

Prerequisites include the Claude Code CLI, a Chrome browser (manual login required for downloads), and optionally the Zotero desktop app and Python 3 for export scripts.

Highlighted Details

• Offers 10 distinct skills, including keyword and advanced search with filters for source category (SCI/EI/CSSCI/北大核心/CSCD), date, and author.
• Supports journal searching, indexing status checks, impact factor retrieval, and browsing journal tables of contents.
• Enables direct PDF/CAJ downloads and export of citations to Zotero or in GB/T 7714 format.
• An orchestrating agent, cnki-researcher, manages multi-step workflows like search-to-export sequences.

Maintenance & Community

Community engagement is facilitated through a Discord server and WeChat groups. Specific details on maintainers, sponsorships, or project roadmap are not detailed in the README.

Licensing & Compatibility

The project is released under the MIT license, which is permissive and generally suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

Users must manually solve any detected captchas, as the system pauses for user intervention. Additionally, users need to be logged into their CNKI account within the Chrome browser for download functionalities to work correctly. The system relies on a running Chrome instance with remote debugging enabled.