Quick Start Tutorial
Get your first Diverga agent running in 5 minutes.
~5 minutes
1
Prerequisites
- Claude Code installed
- Terminal access
- ~5 minutes of your time
Verify Claude Code is installed:
Terminal
claude --version2
Installation
- Clone Diverga or install via marketplace
- Choose your preferred installation method
Install Diverga:
Terminal
# Option 1: Clone from GitHub
git clone https://github.com/yourusername/Diverga.git ~/.claude/plugins/diverga
# Option 2: Marketplace (coming soon)
claude marketplace install diverga3
First Skill - /diverga-help
- Run the help command to see all available skills
- Understand the command output structure
- Familiarize yourself with skill categories
Run your first command:
Terminal
/diverga-helpExpected output:
Diverga v9.0 - Research Coordinator Available Skills: • /diverga-setup - Initial configuration • /diverga-a1 - Research Question Refiner • /diverga-a2 - Theoretical Framework Designer ... Agent Categories: A: Theory & Design (5 agents) B: Literature & Evidence (4 agents) C: Methodology & Analysis (7 agents) ...
4
First Agent - A1 Research Question Refiner
- Trigger the agent with a research question
- Observe how VS methodology presents alternatives
- See the agent's structured response
Start a conversation:
Terminal
I want to study how AI impacts learning outcomes in higher education.Expected output:
[A1-ResearchQuestionRefiner activated] 🔴 CP_RQ_REFINEMENT (REQUIRED) I've analyzed your research interest. Here are three refined research questions across the typicality spectrum: [0] T=0.85 - Modal (Most Common) "How do AI-powered tutoring systems affect student academic performance in higher education?" [1] T=0.55 - Emerging "What is the relationship between AI-mediated personalized learning and metacognitive skill development in undergraduate students?" [2] T=0.25 - Novel "How do human-AI collaborative learning environments influence creativity and critical thinking among graduate students?" Which direction aligns with your research goals?
5
Understanding Checkpoints
- Checkpoints pause for human decision-making
- Three types: REQUIRED, RECOMMENDED, OPTIONAL
- You control critical research decisions
Respond to the checkpoint:
Terminal
approve option 1Checkpoint Types: 🔴 REQUIRED - Must respond before continuing 🟠 RECOMMENDED - Should respond for best results 🔵 OPTIONAL - Can skip or respond Your responses: • "approve option X" - Select option X • "skip" - Skip optional checkpoints • "explain" - Get more details about options
6
Next Steps
- Explore the Memory System for context persistence
- Learn about VS Methodology and T-Scores
- Browse all 24 agent categories
Congratulations! 🎉
You've successfully run your first Diverga agent. You're now ready to leverage all 24 specialized research agents and the full power of the VS methodology.