⚠️ Disclaimer. This repo grows on the fly out of my own thoughts and needs around AI assistance in computational research. It is a working draft, in need of human revision and selective improvement. Helpers and testers welcome!
A Wolfram-centric Claude plugin for AI-assisted computational research. Available in the WolframInstitute marketplace.
- 📁 Turns a folder of resources — code, PDFs, Markdown, notebooks — into an organized git repo, and maintains it.
- 🐺 Imports and exports Wolfram notebooks via Markdown.
- 📚 Grows and maintains a wiki knowledge base.
- 🔍 Gathers and summarizes resources, keeping a Markdown summary and recovery instructions.
- 📦 Converts code into a paclet and builds, documents, and deploys it.
- 📓 Generates expository Wolfram notebooks and publishes them on Wolfram Cloud.
- 📝 Scaffolds a LaTeX or Typst paper and edits the user-owned document on request.
- 🧬 Optionally records the prompt and intent behind every generated artifact.
- 📔 Optionally keeps a running scientific journal in LaTeX or Typst.
- 🧭 Offers a guided tour through the project, and a revision protocol for deliverables.
- ✅ Tracks plans, todos, and state, and can work an opted-in item unattended onto a branch for review.
Distributed through the WolframInstitute plugin marketplace. In Claude Code (CLI / VS Code extension) — the author's setup:
claude plugin marketplace add WolframInstitute/ClaudePluginMarketplace
claude plugin install computational-research@WolframInstituteIn the Claude Desktop app, install from the marketplace GUI.
The plugin works best with Wolfram Engine (or Mathematica), and draws on these MCP servers:
| Server | Required | Purpose | Source |
|---|---|---|---|
| Wolfram (official) | yes | Evaluation, notebook I/O, docs search, tests | Wolfram/AgentTools |
| arxiv-latex-mcp | recommended | Download LaTeX source of arXiv papers | takashiishida/arxiv-latex-mcp |
| arxiv | recommended | Search and download arXiv papers | blazickjp/arxiv-mcp-server |
Install the official Wolfram server from a Wolfram session:
InstallMCPServer["ClaudeCode", "WolframLanguage"]🔑 License seats. Every running kernel — each Wolfram MCP server, each open front-end, each wolframscript call — takes one of your $MaxLicenseProcesses seats. The plugin is MCP-first and checks headroom before spawning a kernel; see the kernel execution policy.
Notes
- Operation in Cowork mode and Chat mode has not been tested.
- On older Wolfram versions the legacy Wolfram/MCPServer paclet still works as a fallback.
- The unofficial sw1sh/WolframMCP server is optional; it adds Wolfram Language LSP support, similar to Serena.
- Running both Wolfram MCP servers at once uses two license seats —
/computational-research:check-envreports live headroom and flags this.
Each skill is invoked by the slash command of the same name, /computational-research:<skill>.
Scripts, templates, project types, and the repo layout are in ARCHITECTURE.md.
| Skill / Command | Description |
|---|---|
| new-project | Scaffold a new project (research, math, paclet-dev, paclet) |
| scaffold-paper | Scaffold a LaTeX or Typst paper, then edit it on request, to the shared writing guide |
| journal | Keep an optional cited LaTeX/Typst journal (def/thm/rem), and take what a paper cannot carry; off by default |
| init-wiki | Create a markdown knowledge base (Wiki/) |
| update-wiki | Update wiki articles, index, and backlinks |
| check-wiki | Audit the wiki for staleness and gaps |
| search-wolfram | Search Wolfram docs, Function Repository, Community, writings |
| search-math | Search MathWorld, nLab, OEIS, DLMF, Wikipedia math |
| add-resource | Add a paper, repo, or page with recovery info |
| cite | BibTeX from an arXiv ID or DOI |
| new-notebook | Build Wolfram notebooks from Markdown (dual-engine: auto-detects a richer converter for frontmatter/LaTeX-math sources) |
| research-notebook | A mathematics paper as a notebook: settled results with complete proofs, experiments quarantined in a Ruliology section, everything numbered and cross-referenced by the front end |
| lean | Drive a Lean/Mathlib formalization session |
| paclet-docs | Generate a symbol reference page per exported paclet function |
| build-paclet | Build a paclet and install it locally |
| publish-paclet | Build with docs, install, publish to the Cloud, deploy the doc pages publicly |
| work | Manage multi-session work items (spec, tasks, hand-off, decisions, progress) |
| next-session | Run the next task in a fresh session, then stop |
| provenance | Track the prompt behind each generated artifact |
| start-tour | Run a guided tour of the project |
| revise | Human revision protocol for deliverables — skill only, no command |
check-env |
Check kernel and MCP availability — command only, no skill |
load-project |
Summarize project status — command only, no skill |
auto-run |
Work an opted-in item unattended, one cold session per task, onto auto/<Item> for a human to review and merge — command only, no skill |
MIT