Community-Driven Development
An AI-augmented, human-in-the-loop workflow that puts quality and safety first.
How it works
The Gnome Starter Pack uses a unique development workflow that combines AI assistance with human oversight. Every step is transparent, traceable, and requires maintainer approval before moving forward.
The Workflow in Detail
1. Issue Creation
Anyone can contribute by opening an issue on GitLab. Describe your feature request, bug report, or improvement idea — no coding required. The more detail you provide, the better, but our AI will ask follow-up questions if needed.
2. AI Triage
Our AI agent analyzes the issue, classifies it (bug, feature, enhancement), and engages in a Q&A loop with the reporter to gather all necessary details. It checks feasibility, estimates complexity, and identifies potential conflicts or dependencies.
3. Action Plan
The AI creates a detailed implementation plan: technical approach, files to modify, testing strategy, and potential risks. This plan is posted to the issue and enters an approval loop — the maintainer can request changes, ask questions, or approve it.
Maintainer Approval
Human oversight is mandatory. The maintainer reviews the action plan before any code is written. This ensures quality, security, and alignment with project goals. Approval triggers the next stage; feedback loops back to refinement.
4. Implementation
Only after approval, the AI branches from test, writes the code, runs tests, and opens a merge request. All changes are transparent and traceable. Beta builds are automatically generated for testing.
5. Review
The maintainer reviews the merge request, tests beta builds, and provides feedback. This is another human-in-the-loop checkpoint. The AI can iterate based on feedback until the code meets quality standards.
6. Merge & Deploy
Once approved, the merge request is merged into main. Our CI/CD pipeline builds Flatpak, DEB, and AppImage packages, runs final tests, and publishes to our repositories. Users receive updates automatically.
Why This Works
Quality
Every change is reviewed by a human maintainer. AI assists, but humans decide. This ensures high code quality and prevents regressions.
Speed
AI handles the repetitive work: analysis, planning, coding, testing. This frees maintainers to focus on decisions and oversight, accelerating development.
Safety
Multiple human checkpoints prevent unwanted changes. The AI cannot merge code without explicit approval. All actions are logged and traceable.
Traceability
Every decision is documented in GitLab issues and merge requests. You can see exactly why and how each change was made.
Accessibility
Non-technical users can contribute ideas without writing code. The AI bridges the gap between user needs and implementation.
Testing
Beta builds are generated automatically for every merge request. Users can test changes before they reach stable, ensuring real-world validation.
Get Started
Ready to contribute? Open an issue, request a feature, or report a bug on GitLab. Our AI will guide you through the rest.