Alien Marketplace
The Alien Marketplace is a plugin registry for Claude Code that extends AI agents with specialized research and data tools. Plugins are installed directly into Claude Code and provide MCP servers, skills, and scenario-driven workflows — giving Claude domain expertise it does not have out of the box.
The marketplace is open source: github.com/the-alien-club/claude-marketplace
Setup
Claude Code
Add the marketplace, install a plugin, and reload — three commands:
/plugin marketplace add the-alien-club/claude-marketplace#main
/plugin install openscience@alien
/reload-plugins
That's it. Claude Code downloads the plugin definition, registers its MCP servers, and makes its skills available. The plugin is ready to use immediately.
Claude Cowork
Claude Cowork support is planned but not yet available. This section will be updated when Cowork integration is ready.
How It Works
The marketplace uses Claude Code's plugin system. Each plugin can bundle:
- MCP servers — Expose structured tools that Claude can call (search, retrieve, analyze).
- Skills — Scenario-driven guides that teach Claude how to approach domain-specific tasks (e.g., conducting a literature review, assessing research impact).
- Configuration — Pre-wired server URLs and authentication so setup is minimal.
When you enable a plugin, Claude Code registers its MCP servers and loads its skills. Claude then has access to the plugin's tools and knows when and how to use them.
Available Plugins
| Plugin | Category | Description |
|---|---|---|
| openscience | Research | OpenAIRE Research Graph + bioRxiv/medRxiv preprint data clusters |
OpenAIRE — Open Science Toolkit
The openscience plugin connects Claude to the OpenAIRE Research Graph, one of the largest open metadata aggregations in scholarly communication — 600M+ research products including publications, datasets, software, and other research outputs across all disciplines.
It also provides access to bioRxiv and medRxiv preprint data clusters for life sciences and medical research.
MCP Servers
The plugin registers three MCP servers:
| Server | Endpoint | Data Source |
|---|---|---|
openaire-beta | openaire.mcp.beta.alien.club | OpenAIRE Research Graph — publications, datasets, projects, authors, organizations |
datacluster-biorxiv-beta | biorxiv.mcp.beta.alien.club | bioRxiv preprint corpus |
datacluster-medrxiv-beta | medrxiv.mcp.beta.alien.club | medRxiv preprint corpus |
OpenAIRE Tools
The OpenAIRE MCP server exposes ~28 tools across four API layers:
| Layer | Capabilities |
|---|---|
| Graph API v1 | Organizations, projects, persons, data sources |
| Graph API v2 | Research products, bibliometric filters, datasets |
| ScholeXplorer | Cross-product relationships (citations, supplements, references) |
| Composite / Analytical | Citation networks, co-authorship analysis, trend analysis, author profiles |
Key tools include:
search_research_products— Search publications, datasets, and software by keyword, author, field of science, or bibliometric class.get_research_product_details— Retrieve full metadata for a DOI including all four bibliometric indicators.find_by_influence_class/find_by_popularity_class— Find landmark or trending papers using OpenAIRE's normalized bibliometric rankings (C1 = top 0.01% through C5 = bottom 90%).get_citation_network— Build citation graphs from seed DOIs.analyze_coauthorship_network— Map research collaboration patterns.analyze_research_trends— Track publication volume over time for a topic.search_datasets/search_projects— Discover research datasets and funded projects.
Research Scenarios
The plugin includes a skill (explore-openaire) with nine pre-built research scenarios. When Claude detects that your question matches a scenario, it follows a structured, multi-step workflow using the right tools in the right order:
| Scenario | Use Case |
|---|---|
| Literature review | Survey a field, find key papers, trace citation lineage |
| Author landscape | Explore a researcher's work and collaborations |
| Project impact | Assess outputs of a funded research project |
| Cross-domain discovery | Find methods or data from outside your home field |
| Dataset discovery | Find research datasets related to a topic |
| Co-citation analysis | Reveal intellectual structure via citation patterns |
| Bibliometric assessment | Find landmark, trending, or high-impact papers |
| Find primary publication | Locate the canonical paper for a tool, method, or dataset |
| Assess dataset relevance | Evaluate whether a dataset is suitable for your research |
Bibliometric Classes
OpenAIRE assigns four independent bibliometric indicators to research products. Each is ranked into classes:
| Class | Percentile | Meaning |
|---|---|---|
| C1 | Top 0.01% | Landmark / foundational |
| C2 | Top 0.1% | Highly notable |
| C3 | Top 1% | Field-leading |
| C4 | Top 10% | Above average |
| C5 | Bottom 90% | Average |
The four metrics — influence (field-normalized impact), popularity (recent citation velocity), impulse (early momentum), and citation count (absolute) — can be combined to characterize a paper's significance. For example, C1 influence + C1 popularity identifies an established paper that is still actively cited.
bioRxiv and medRxiv Data Clusters
The preprint data clusters use the same datacluster_* tool interface as your own data clusters (see AI Agent Integration). This means Claude can search preprints using the same tools it uses for your private documents:
datacluster_list_datasets— List available preprint collectionsdatacluster_keyword_search— Full-text search across preprintsdatacluster_vector_search_chunks— Semantic similarity searchdatacluster_get_entry_content— Read full preprint text
Example Usage
Once the plugin is enabled, you can ask Claude research questions directly:
"Find the most influential papers on transformer architectures in NLP published since 2017, and summarize the top 3."
Claude will use the OpenAIRE tools to search by influence class, retrieve details for the top results, and synthesize a summary — all grounded in real metadata from the Research Graph.
"Search bioRxiv for recent preprints on CRISPR base editing in crops and compare their approaches."
Claude will search the bioRxiv data cluster, read the relevant preprints, and provide a comparative analysis.
Next Steps
- AI Agent Integration — How MCP servers connect AI agents to your data
- Search and Query — The search capabilities underlying MCP tools
- Deploy MCP — Deploy MCP servers on your own data clusters