This project provides a comprehensive suite of 433 academic skills and 34 tools for AI agents, designed to equip them with a "research brain." It targets AI agents and users seeking to automate complex academic workflows, transforming conversational AI into capable research assistants that can perform tasks like literature reviews, data analysis, and grant writing. The primary benefit is enabling AI agents to execute practical, research-oriented tasks beyond simple information retrieval.

How It Works

The core approach integrates structured, step-by-step "SKILL.md" guides with direct API integrations to 18 free academic databases. This combination allows AI agents to not only understand research concepts but also to actively perform actions such as searching databases (OpenAlex, PubMed, arXiv), resolving DOIs, tracking citations, and retrieving full-text articles. The skills cover the entire research lifecycle, from literature discovery and systematic reviews to data analysis (including causal inference) and academic writing. This architecture bridges the gap between LLM capabilities and the rigorous, procedural demands of academic research, enabling agents to function as practical research assistants.

Quick Start & Requirements

• Research-Claw / OpenClaw Users: Install via openclaw plugins install @wentorai/research-plugins. Plugins are automatically loaded after restarting the gateway.
• Other 40+ Agent Frameworks (e.g., Claude Code, Cursor): Install skills only using npx skills add wentorai/research-plugins. This method adds SKILL.md files but not the API tools.
• Prerequisites: Requires a compatible AI agent framework. No specific hardware or software versions are mandated beyond the framework's requirements.
• Links: Research-Claw

Highlighted Details

• 433 structured academic skills covering literature, research methods, data analysis, academic writing, specific domains, and efficiency tools.
• 34 API tools provide direct access to 18 free academic databases and services like OpenAlex, CrossRef, PubMed, arXiv, and Unpaywall.
• Skills employ progressive loading (categories -> sub-categories -> individual skills) to manage context efficiently.
• Includes 6 curated lists of high-quality resources for discovery and reference.

Maintenance & Community

Contributions for new skills and resources are welcomed, with guidance provided in CONTRIBUTING.md. The project relies on publicly available academic resources and adheres to open-source principles for its original content.

Licensing & Compatibility

All original content within this project is released under the MIT license, which permits commercial use and integration into closed-source projects. Third-party components and data sources retain their respective licenses.

Limitations & Caveats

Users employing pnpm must use the openclaw plugins install command; loading plugins directly from node_modules via plugins.load.paths is blocked due to security checks related to pnpm's hard-linking mechanism. Furthermore, frameworks other than Research-Claw/OpenClaw will only have access to the SKILL.md files, not the integrated API tools, when using the npx skills add method.