CNKI Skills
CNKI Skills for Claude Code
A Claude Code skill set that drives CNKI through Chrome DevTools MCP: search papers, browse journals, download PDFs and export to Zotero
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Drives the browser and runs scripts on CNKI pages
- Triggers file downloads and sends records to an external Zotero
Install
Manual install
claude mcp add chrome-devtools -- npx -y chrome-devtools-mcp@latestThe Chrome DevTools MCP server; the skills do not work without it.
This is third-party code. Review the repository files before installing.
What it does
The set gives Claude Code ten skills and an orchestrating agent for the Chinese academic database CNKI. The skills run keyword and advanced search with filters by author, journal, date and source category, parse the results page, paginate, pull full paper metadata, look up journals and their indexing, open issue tables of contents, trigger PDF or CAJ downloads and format a GB/T 7714 citation or push a record to Zotero. Everything runs through asynchronous calls in Chrome DevTools MCP, without screenshot parsing. Downloads use the user's own logged-in browser session; the skill does not solve captchas itself but pauses and asks the user to solve them manually.
Who it is for. For researchers, teachers, graduate students and technical writers who work with Chinese academic literature in CNKI.
Good fit when
- You need to find papers in CNKI and get a structured result list
- You need to check a journal, its indexing and metrics or open an issue's table of contents
- You need to download a paper PDF and push the record into Zotero with a ready citation
Not a fit when
- You have no CNKI access with download rights of your own
- You need a non-Chinese database; the skills are built around the CNKI layout
Example request
Find CSSCI papers on machine learning in CNKI for 2020-2025 and export the first five to ZoteroLimitations
You need the Chrome DevTools MCP server installed and Chrome running with a debugging port. Downloads work only when you are logged in to CNKI under an account with download rights, usually institutional or paid; the skills do not bypass passwords or subscriptions. Captchas are solved by the user manually. The Zotero export needs Python 3 and the Zotero desktop app running. Selectors are tied to the current CNKI markup and can break when it changes.
How to disable. Remove the skills/cnki-* folders and agents/cnki-researcher.md from your project's .claude directory or from the global ~/.claude. Optionally remove the chrome-devtools server from your MCP config.
Security check
- Drives the browser and runs scripts on CNKI pages
- Triggers file downloads and sends records to an external Zotero
README in short
The README, in English and Chinese, describes ten skills and the cnki-researcher agent that chains them into flows such as search, filter and export to Zotero. Installation adds the chrome-devtools MCP server, copies the skills/ and agents/ folders into the .claude directory and starts Chrome with a debugging port. Each skill runs in one or two calls: navigate plus one asynchronous evaluate_script, with no screenshots or OCR. The agent detects captchas and pauses for the user to solve them manually. The push_to_zotero.py script sends records to the Zotero desktop app through its connector.
SKILL.md
--- name: cnki-search description: Search CNKI (中国知网) for papers by keyword. Use when the user wants to find academic papers on a topic. argument-hint: "[search keywords]" --- # CNKI Basic Search Search CNKI for papers using keyword(s). Returns result count and structured result list (titles, URLs, authors, journal, date) in a single call. ## Arguments $ARGUMENTS contains the search keyword(s) in Chinese or English. ## Steps ### 1. Navigate Use `mcp__chrome-devtools__navigate_page` -> `https://kns.cnki.net/kns8s/search` ### 2. Search + extract results (single evaluate_script, NO wait_for)
FAQ
Do the skills bypass paid access to CNKI?
No. Search and browsing are open, and downloads go through the site's own download buttons in your logged-in session, so you need an account with download rights. The skills do not bypass subscriptions or captchas.
Why is Chrome DevTools MCP required?
It lets the agent control Chrome tabs and run scripts on CNKI pages. Without that server the skills do not work.
Related
CLI, skill and Python library for browser control: the agent clicks, fills forms and reads pages over CDP
A skill that gathers the last 30 days of discussion on a topic from Reddit, X, YouTube, HN, Polymarket and GitHub into one brief
The Chrome DevTools team's official MCP server: the agent drives a live Chrome, reads network and console, and records performance traces
A fast Rust CLI for agent browser automation: accessibility snapshots with element refs, an MCP server and skills