Google Scholar Skills for Claude Code is a set of tools designed to enable programmatic interaction with Google Scholar directly from the Claude Code command-line interface. It targets users who leverage Claude Code for research and academic tasks, providing capabilities for searching papers, tracking citations, obtaining full-text access links, and exporting references to Zotero, thereby streamlining academic workflows.

How It Works

The project utilizes the Chrome DevTools Protocol (CDP) via the chrome-devtools-mcp library to interact with a running Chrome browser instance. All data extraction is performed through DOM scraping using CSS selectors, as Google Scholar lacks a public API. Each skill operates via a single asynchronous evaluate_script call, ensuring fast and reliable execution without relying on brittle screenshot parsing or OCR. A key design choice is the use of data-cid (cluster ID) as a primary identifier across skills for citation tracking and cross-referencing, and BibTeX data is retrieved via navigate_page to circumvent CORS restrictions. The system includes CAPTCHA detection, pausing execution for manual user intervention when necessary.

Quick Start & Requirements

• Prerequisites: Claude Code CLI installed, Chrome browser with remote debugging enabled (--remote-debugging-port=9222), Zotero desktop app (optional for export), Python 3 (optional for Zotero push script).
• Installation:
  1. Install Chrome DevTools MCP server: claude mcp add chrome-devtools -- npx -y chrome-devtools-mcp@latest
  2. Clone the repository and copy skills/agents to .claude/ directory, or integrate into an existing project.
  3. Start Chrome with remote debugging enabled.
  4. Launch Claude Code; skills are automatically detected.
• Documentation/Demo: Links to English and Chinese documentation, WeChat Official Account, WeChat Group, and Discord are provided.

Highlighted Details

• Skills: Offers gs-search, gs-advanced-search (by author, date, etc.), gs-cited-by, gs-fulltext (PDF, DOI, Sci-Hub links), gs-navigate-pages, and gs-export to Zotero.
• Agent: The gs-researcher agent orchestrates all skills, supporting multi-step workflows and CAPTCHA handling.
• Data Extraction: Pure DOM scraping via CSS selectors ensures direct access to scholar.google.com data.
• Zotero Integration: Direct export to Zotero via BibTeX extraction, with an optional Python script for pushing.

Maintenance & Community

The project provides links to a Discord server and WeChat groups for community interaction and support, indicating an active, albeit potentially bilingual, community.

Licensing & Compatibility

The project is released under the MIT License, which is permissive and generally compatible with commercial use and closed-source projects.

Limitations & Caveats

The reliance on DOM scraping makes the project susceptible to breakage if Google Scholar's website structure changes. CAPTCHA detection requires manual user intervention, potentially interrupting automated workflows.