7a336e6e/Configuring CI/CD

Configuring CI/CD

Goal

Automate the verification and delivery process so that code is always tested, secure, and ready for deployment.

When to Use

• Starting a new project (set up the pipeline immediately).
• Adding a new quality check (linting, type checking).
• Automating deployment to staging/production.

Instructions

1. Define the Workflow

Create .github/workflows/ci.yml.

• Triggers: push to main, pull_request to main.
• Jobs: Split into lint, test, build.

2. Lint & Type Check (Fastest)

Fail fast.

• Run eslint, prettier, black, ruff.
• Run type checkers: tsc, mypy.

3. Test (The Gatekeeper)

Run unit and integration tests.

• Set up services (Postgres, Redis) using services in GitHub Actions.
• Run pytest or vitest.
• Upload coverage reports.

4. Build & Deploy (Production)

Only run on push to main.

• Build Docker images.
• Push to registry (GHCR/ECR/Docker Hub).
• Trigger deployment (e.g., via webhook or SSH).

Example Workflow Snippet

name: CI
on: [push, pull_request]
jobs:
  test:
    runs-on: ubuntu-latest
    services:
      postgres:
        image: postgres:15
        env:
          POSTGRES_PASSWORD: password
        ports:
          - 5432:5432
    steps:
      - uses: actions/checkout@v3
      - uses: actions/setup-python@v4
      - run: pip install -r requirements.txt
      - run: pytest

Constraints

✅ Do

• DO: Cache dependencies (actions/cache) to speed up builds.
• DO: Separate "Checks" (PRs) from "Deployments" (Main).
• DO: Use secrets for all sensitive config (${{ secrets.DB_URL }}).
• DO: Pin action versions (actions/checkout@v3).

❌ Don't

• DON'T: Hardcode credentials in the workflow file.
• DON'T: Skip tests on the main branch deployment.
• DON'T: Allow the pipeline to succeed if the linter fails.

Output Format

• .github/workflows/*.yml files.

Dependencies

• shared/environment-config/SKILL.md
• security/auditing-dependencies/SKILL.md