academic-research-skills: Instructions file for Claude Code

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

academic-research-skills is a plugin marketplace for AI-powered academic research tools. It provides 58+ skills across 8 academic platforms (CNKI, Google Scholar, ScienceDirect, Web of Science, PubMed, IEEE Xplore, Nature, Zotero) for Claude Code, Copilot CLI, Codex, Gemini CLI, and other AI assistants.

Repository Structure

academic-research-skills/
├── plugins/                 # One directory per plugin
│   └── {platform}-{action}/
│       ├── .claude-plugin/
│       │   └── plugin.json  # Plugin metadata (required)
│       └── skills/
│           └── {skill-name}/
│               ├── SKILL.md     # Skill instructions (required)
│               ├── scripts/     # Executable scripts (optional)
│               └── references/  # Reference documents (optional)
├── scripts/                 # Validation tools
│   ├── validate_skills.py   # SKILL.md validator
│   └── validate_plugin_json.py
└── docs/
    └── architecture.md      # System architecture

Development Commands

Validation (run before committing)

python scripts/validate_skills.py      # Validate all SKILL.md files
python scripts/validate_plugin_json.py # Validate plugin.json files

Linting

ruff check plugins/ scripts/           # Lint Python files
ruff format --check plugins/ scripts/  # Format check

Plugin Architecture

Two Plugin Types

1. Browser-based skills (CNKI, GS, SD, WoS, PubMed, IEEE): Use Chrome DevTools MCP for DOM automation. Each plugin.json includes mcpServers config for chrome-devtools-mcp.

2. Pure prompt-based skills (Nature): Work with any LLM, no browser required.

Plugin Naming Convention

Pattern: {platform}-{action}

Platform	Prefix
CNKI	cnki-
Google Scholar	gs-
ScienceDirect	sd-
Web of Science	wos-
PubMed	pm-
IEEE Xplore	ieee-
Nature	nature-
Zotero	zotero-