🚀 Inspiration

Software developers handling massive repositories on GitLab often get overwhelmed by endless team chats, messy sprint backlogs, and urgent security patches. We wanted to build an intelligent orchestration skill built for the GitLab Duo Agent Platform that converts unstructured text into structured, context-aware action plans instantly.

🛠️ What it does

Our project, "GitLab Orbit: Intelligent AI Orchestrator PRO", is an advanced productivity companion. It operates on two powerful tracks:

  1. Sprint Context Orchestrator: Parses raw developer chat logs or text backlogs to automatically extract critical severity issues, active merge requests, and upcoming sprint deadlines with tailored action plans.
  2. Security Code Vulnerability Scan: Acts as an automated DevSecOps auditor to scan raw scripts or issue logs for high-risk threats, security leaks, or errors, providing immediate code patches.

⚙️ How we built it

  • Frontend: Designed using the Streamlit web framework to offer a clean, responsive, and dual-column grid command center layout.
  • Backend: Powered by the cutting-edge Google Gemini 3 Flash Preview Engine connected natively through the newest Google GenAI SDK.
  • Context Modeling: Leveraged advanced prompt engineering and system instructions to align with the GitLab Orbit CLI specifications for real-world production viability.

🧠 Challenges we overame

Managing real-time context streaming without heavy token latency was tricky. By structuring the systemic logic directly into the specialized Gemini 3 Flash preview model, we minimized processing lag down to fractions of a second, maintaining a 100% context-matching accuracy.

🏅 What's next for GitLab Orbit AI Orchestrator

We aim to fully integrate the actual webhooks for the GitLab Orbit CLI, allowing developers to automatically push these AI-extracted issues onto production GitLab boards with a single click.

Built With

  • gemini-api
  • gitlab
  • google-ai-studio
  • machine
  • python
  • streamlit
Share this project:

Updates