research

InfAU Skill-Base

Academic skills for Claude Code at the InfAU chair (Informatik in der Architektur und Urbanistik), Bauhaus-Universität Weimar.

Installation (Private Repository)

Prerequisites

Authenticate with GitHub:

gh auth login

Add the Marketplace

/plugin marketplace add Bauhaus-InfAU/infau-skill-base

Install Plugins

/plugin install research@infau-skills
/plugin install docs@infau-skills
/plugin install grasshopper@infau-skills
/plugin install monet@infau-skills

For Auto-Updates (Optional)

Set GitHub token in your shell config (~/.bashrc or ~/.zshrc):

export GITHUB_TOKEN=ghp_xxxxxxxxxxxxxxxxxxxx

What is this?

This repository contains domain-specific plugins for Claude Code that help with common academic tasks. Each domain plugin includes skills (reusable prompts) and commands (slash commands) that guide Claude through specific workflows.

Available Domains

Research (research/)

Research planning and problem selection framework for academic research in computational design, urban planning, and sustainable building.

Commands:

• /research-plan - Plan research projects using scientific problem selection framework
• /research-question - Improve research questions using FINER/PICO frameworks
• /deep-research - Conduct comprehensive multi-source research with citation tracking

Example:

/research-question How can computational simulation evaluate the impact of micro-scale interventions in fostering spatial synchronization?

Claude will:

1. Diagnose specific problems with your question
2. Select an appropriate framework (FINER, PICO, or InfAU-specific)
3. Ask clarifying questions one at a time
4. Show before/after comparisons as you improve

Grasshopper (grasshopper/)

Tools for working with Grasshopper (Rhino) parametric definitions.

Commands:

• /ghx-to-llm - Convert GHX files to LLM-readable markdown with cluster resolution and letter-based cross-referencing

Monet (monet/)

AI image generation for architecture using Google Gemini.

Commands:

• /monet - Generate architecture-focused images with structured prompts and style references

Example:

/monet new

Claude will:

1. Scaffold a monet/ directory with context, references, prompts, and results folders
2. Read your project context and style references
3. Author structured prompts (JSON or narrative)
4. Generate images via Gemini and auto-version results

Docs (docs/)

Document processing and conversion workflows.

Commands:

• /pdf-to-md - Convert PDF files to Markdown using Docling

Structure

skill-base/
├── .claude-plugin/                   # Root plugin manifest + marketplace
│   ├── plugin.json
│   └── marketplace.json
├── research/                         # Research domain plugin
│   ├── .claude-plugin/
│   ├── commands/
│   │   ├── research-plan.md
│   │   ├── research-question.md
│   │   └── deep-research.md
│   └── skills/
│       ├── research-question/
│       ├── scientific-problem-selection/
│       └── deep-research/
├── grasshopper/                      # Grasshopper domain plugin
│   ├── .claude-plugin/
│   ├── commands/
│   │   └── ghx-to-llm.md
│   └── skills/
│       └── ghx-to-llm/
├── monet/                            # AI image generation plugin
│   ├── .claude-plugin/
│   ├── commands/
│   │   └── monet.md
│   └── skills/
│       └── monet/
├── docs/                             # Document processing plugin
│   ├── .claude-plugin/
│   ├── commands/
│   │   └── pdf-to-md.md
│   └── skills/
│       ├── pdf/
│       └── pdf-to-md/
├── _templates/                       # Templates for new skills
├── CLAUDE.md
├── CONTRIBUTING.md
└── README.md

Contributing

Want to add your own skill? See CONTRIBUTING.md for a step-by-step guide. No programming experience required.

License

MIT License - see LICENSE for details.