When building CI/CD pipelines, keep them scope-focused and deterministic, and add explicit quality gates so failures stop the deployment. Apply: - Put workflows in the conventional location and naming your team standard expects (e.g., `.github/workflows/main.yml`). - Keep the CI/CD goal minimal (avoid extra tooling/framework bootstraps when the purpose is learning/automation). - Run formatting/validation before deployment (e.g., `terraform fmt` + `terraform validate`), and structure steps in a clear order (`init → plan → apply`-style). - Ensure any step error fails the workflow and prevents deploy (default GitHub Actions behavior already stops on non-zero exit codes). Example (GitHub Actions): ```yaml name: CI/CD on: [push] jobs: deploy: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - name: Build/Test run: | echo "run tests/build here" - name: Validate (fail fast) run: | terraform fmt -check terraform validate - name: Deploy if: success() run: | terraform init terraform plan -out=tfplan terraform apply -auto-approve tfplan ```