Inspiration

During software development, developers can overlook missing tests, outdated documentation, or potential risks when preparing to merge code. I wanted to explore how AI-powered automation could make merge request reviews more consistent and help developers identify issues before merging.

What It Does

Merge Readiness Coach is a GitLab Duo flow that reviews merge requests across three areas:

• Test Coverage: Identifies potential gaps in testing and suggests improvements.
• Documentation: Checks for documentation issues related to code changes.
• Risk Assessment: Provides an overall risk level and a readiness verdict.

The flow posts its findings directly in the merge request, helping developers understand what needs attention before merging.

How I Built It

I built the project using GitLab's Duo Agent Platform and a GitLab repository containing Python source code, automated tests, documentation, and a CI/CD pipeline.

The flow uses GitLab tools to inspect merge request changes and relevant repository files, then creates a review comment with its findings. I also used GitLab CI/CD to run the project's tests and verify the changes.

Challenges I Faced

One of my biggest challenges was learning GitLab's interface and understanding how AI flows, merge requests, and CI/CD pipelines work together. I also had to troubleshoot configuration issues and improve the code and documentation based on the flow's feedback.

Working through these challenges helped me better understand how to connect AI-assisted workflows with real software development processes.

What I Learned

This project gave me practical experience with GitLab Duo Agent Platform, AI-powered development workflows, merge request reviews, and CI/CD. It also taught me the importance of testing, clear documentation, and verifying results instead of relying on AI feedback alone.

What's Next

Future improvements could include deeper test analysis, additional code quality checks, and broader support for identifying potential issues before code is merged.

Built With

• ai
• automation
• ci/cd
• gitlab
• python
• testing