Gemini for DevOps Engineers

For DevOps teams, one of the biggest priorities is ensuring that every code change is clean, secure, and free from issues that could trigger an outage — whether it's in development, staging, or production. On top of that, security vulnerabilities in the code can lead to serious risks, and no team wants to deal with that headache.

This is where Gemini AI steps in to make life easier. By integrating Gemini into your GitHub Actions workflow, pull requests (PRs) get an extra layer of review before any code gets deployed. Instead of relying solely on manual review, Gemini checks for potential impacts, security vulnerabilities, and logical issues. If everything passes Gemini’s review, the DevOps team can confidently move forward with the deployment.

In this blog, we'll walk you through how Gemini AI can review pull requests using GitHub Actions and how it helps DevOps teams ship code with confidence — faster, safer, and with fewer sleepless nights.

Why Gemini AI for Code Review?

Gemini offers several key benefits for code review:

Reduced Manual Effort: Gemini automates the initial code review process, freeing up developers to focus on more complex tasks.

Improved Code Quality: Gemini can identify potential bugs, security vulnerabilities, and deviations from coding standards, leading to more robust and maintainable code.

Enhanced Consistency: Gemini helps enforce consistent coding practices across your team, improving collaboration and reducing errors.

Faster Feedback Loops: Get instant feedback on your code changes, accelerating the development cycle and enabling faster iteration.

Code Review by Gemini AI

For this newsletter, we will create a customized GitHub Action workflow using Python that will scan the pull request code using Gemini 1.5 Pro. It will then provide a code review and suggest improvements.

The workflow will deliver a full report in a table format, outlining the code and the recommended improvements.

Setting up Gemini Code Review in GitHub Actions

1. To use this GitHub Action, you need an Gemini API key. If you don't have one, sign up for an API key at Google AI Studio.
2. Add the Gemini API key as a GitHub Secret in your repository with the name GEMINI_API_KEY. You can find more information about GitHub Secrets here.
3. Create a `.github/workflows/code-review.yml` file in your repository and add the following content:

name: Code Scan

on:
  pull_request:
    types: [opened, synchronize]

permissions: write-all

jobs:
  scan:
    runs-on: ubuntu-latest


    steps:
      - uses: actions/checkout@v3

      - name: Checkout gemini-code-scanner
        uses: actions/checkout@v3
        with:
          repository: GDGCloudLahore/gemini-code-scanner 
          path: gemini-code-scanner 
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} 
          

      - name: Gemini Code Scanner
        uses: ./gemini-code-scanner  
        env:
          GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        
        

4. Click on “add commit” and create a new branch. The PR will automatically be triggered and run the Gemini scanner.

The output will look like the below one.

The Gemini output looks like this on a pull request.

In this example, Gemini identified a potential security vulnerability in the code and provided a detailed explanation along with suggestions for remediation. This immediate feedback allows developers to address the issue early in the development process, preventing it from becoming a larger problem down the line.

How It Works

This GitHub Action uses the Gemini AI API to provide code review feedback. It works by:

1. Analyzing the changes: It grabs the code modifications from your pull request and filters out any files you don't want reviewed.
2. Consulting the Gemini model: It sends chunks of the modified code to the Gemini for analysis.
3. Providing feedback: Gemini AI examines the code and generates review comments.
4. Delivering the review: The Action adds the comments directly to your pull request on GitHub.

Contributions

This newsletter and the Gemini scanner featured within were created by Adil Shahzad . As the tool is still in beta, any feedback or suggestions for improvement are highly valuable. Feel free to contribute by opening a pull request:

gemini-code-scanner repository.