AI Native Notes
Back to Gemini CLI
Gemini CLIAdvanced3 min read

Gemini GitHub Actions — CI Automation and Review

Integrate Gemini CLI into GitHub Actions to build automated code reviews, document parity checks, and AI-powered static analysis workflows.

github-actionsci-cdautomationcode-reviewheadless

Official References: Headless Mode · Continuous Integration

Beyond the Terminal: A Teammate in CI/CD

The true value of Gemini CLI is realized when it is woven into the team's workflow. By combining it with GitHub Actions, you can build systems where AI provides code reviews for every Pull Request (PR) or automatically checks the consistency of complex locale files.

1. Basic Setup and Authentication

Since interactive login is impossible in a CI environment, use the API Key method for authentication.

# .github/workflows/gemini-review.yml
jobs:
  gemini-review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with:
          node-version: '20'
      - name: Install Gemini CLI
        run: npm install -g @google/gemini-cli
      - name: Run Review
        env:
          GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}
        run: |
          # Run in Headless mode for automated review
          gemini --headless "Review the security and performance of changed files. Write the output in PR comment format."

2. Practical Use Cases

A. Locale Parity Checks

Ensures that keys between translation files (e.g., ko.json, en.json) match and identifies missing translations.

gemini --headless "Compare all JSON files in the messages/ folder and report any missing keys."

B. AI-Powered Semantic Code Review

Points out logical errors or readability issues in business logic that simple linters might miss.

gemini --headless "Verify if the modified files in this PR follow the architectural rules defined in our GEMINI.md."

C. Automatic Documentation Proposals

Automatically suggests PRs to update READMEs or API documentation based on code changes.

3. Synergy Between GEMINI.md and CI

Gemini CLI in the CI environment reads the GEMINI.md at the project root just like it does locally. This ensures a consistent experience between local and CI environments.

CI-Specific Instruction Example:

"When running in CI, keep responses as concise as possible; only include specific suggestions when an error is discovered."

4. Security and Cost Optimization

  1. Protect Secrets: Always store the GEMINI_API_KEY in GitHub Secrets and ensure it is not exposed in logs.
  2. Headless Mode: Shorten CI execution time by skipping the interactive interface.
  3. Scoped Context: Limit the range of files the agent reads (e.g., gemini "@src @messages ...") to optimize token usage.

5. Advanced: Combining Static Analysis with AI

Static tools (ESLint, SonarQube) are more powerful when combined with Gemini CLI. Pipe the linter's output into Gemini and have it suggest exactly how to fix the errors.

npm run lint > lint-error.txt
gemini --headless "@lint-error.txt Suggest the safest code to fix these errors."

Next to Read

Connected Guides