Week 1 — Setting up SustainHub for Open Source Community Stability

Hello world!
Welcome to my first blog post documenting my Google Summer of Code (GSoC) 2025 journey with the Orthogonal Research and Education Lab.

---

Week 1: Environment and Agent Setup

This week marked the official beginning of the coding phase, and I focused on building a strong foundation for the SustainHub simulation.

What I accomplished:

1. Environment Initialization

• Built a GitHub-like simulation environment in Python.
• Designed task types:
  • Bug fixes
  • Feature requests
  • Documentation updates
• Set up dynamic task queues, difficulty levels, and task deadlines.

2. Agent Definition and Role Assignment

• Implemented four key agent roles as Python classes:
  • Maintainers: Oversee project quality and merge PRs.
  • Contributors: Tackle bugs and implement features.
  • Innovators: Propose new ideas.
  • Knowledge Curators: Maintain documentation and knowledge artifacts.
• Initialized agent attributes:
  • Expertise Level (Apprentice to Expert)
  • Task Load
  • Success Rate

3. Basic Task Assignment Logic

• Implemented rule-based assignment based on agent availability and expertise.
• Defined parameters to simulate realistic collaboration.

---

Learnings:

• Python’s OOP principles came in handy for modularizing agent behavior.
• Carefully modeling environment variables helped clarify how complexity and deadlines affect task flow.
• Initial testing gave early insights into potential bottlenecks in future task allocation stages.

---

Next Up:

In Week 2, I’ll be working on:

• Implementing Multi-Armed Bandit (MAB) strategies for dynamic task assignment.
• Refining the task generation system to improve diversity and alignment with agent expertise.

Stay tuned as SustainHub evolves into a robust, intelligent agent-based framework for open-source project stability.

---

Thanks to my mentors for their guidance and support throughout this journey so far!
— Vidhi Rohira