ci-cd-and-automation

CI/CD and Automation - Automated Continuous Integration and Deployment

Skill Overview

The CI/CD and Automation skill helps you automatically set up and manage continuous integration and deployment pipelines, ensuring that every code change undergoes a complete quality assurance process, including linting, type checking, unit testing, building, and security auditing. It implements “shift-left” testing and enables fast, secure production deployments.

Use Cases

1. Setting Up a CI Pipeline for a New Project

When launching a new project, this skill can quickly establish a complete quality gate process. From basic linting, type checking, and unit testing to integration testing and E2E testing, ready-made GitHub Actions configuration templates are available for every stage. It is especially suitable for Node.js projects, supporting automated configuration for PostgreSQL database integration tests and Playwright E2E tests.

2. Optimizing CI/CD for Existing Projects

If your CI pipeline runs slowly—taking more than 10 minutes—this skill provides systematic optimization strategies, including dependency caching, parallel jobs, matrix builds, and on-demand execution. It also includes configuration for automated dependency updates (Dependabot/Renovate) and Build Cop role setup, helping teams keep their CI pipelines running smoothly.

3. Deployment Strategies and Rollback Mechanisms

Supports configuration of multiple deployment strategies: preview deployments (automatically deploying a test environment for every PR), feature flags (decoupling deployment from release), staged releases (Staging → Production), and manual rollback workflows. It emphasizes that every deployment should be reversible and provides complete GitHub Actions configuration examples.

Core Features

1. Complete Quality Gate Process

Automatically configures a complete validation pipeline from code commit to production release: Linting → TypeScript type checking → Unit testing → Build → Integration testing → E2E testing (optional) → Security auditing → Bundle size checking. Each stage must pass strictly before proceeding, ensuring that the “shift-left” principle is implemented—problems are discovered during development rather than debugged in production.

2. GitHub Actions Automation Configuration

Provides ready-to-use YAML configuration templates covering basic CI pipelines, database integration tests, E2E tests, preview deployments, manual rollbacks, and other workflows. Supports secure GitHub Secrets management, service containers (PostgreSQL) with health checks, build artifact uploads, and more. All configurations follow best practices, avoid hardcoded secrets, and use caching to improve performance.

3. AI-Powered Feedback Loop

When CI fails, this skill establishes a clear feedback mechanism: CI failure output is sent directly to the AI agent, which automatically fixes lint errors (by running npm run lint --fix), identifies and resolves type errors, follows the debugging skill to address test failures, and checks configuration and dependencies to resolve build errors. It then pushes the fixes and retriggers CI, forming a complete automated repair loop.

Frequently Asked Questions

How do I build a complete CI/CD pipeline from scratch?

Start with a basic GitHub Actions configuration and gradually add quality gates. First, configure .github/workflows/ci.yml with linting, type checking, unit testing, and build steps. Then add integration tests, such as database tests, and E2E tests as needed. Finally, configure branch protection rules to ensure that CI must pass before merging. This skill provides complete configuration templates and best-practice guidelines.

How can I optimize a CI pipeline that runs too slowly?

Apply the following strategies in order of impact: use the caching options in actions/setup-node to cache dependencies; split linting, type checking, testing, and building into parallel jobs; use path filters to skip unrelated jobs, such as skipping E2E tests when only documentation changes; use matrix builds to shard test suites across multiple runners; and remove slow tests from the critical path, running them on a scheduled basis instead. The goal is to complete the full test suite within 10 minutes.

How can I efficiently debug and fix CI failures?

Establish a clear feedback loop: copy the specific CI failure messages and provide them to the AI agent, stating, “The CI pipeline failed with the following errors. Please fix them, verify the fixes locally, and push again.” The AI will automatically apply the appropriate repair strategy based on the error type: run automatic fixes for lint errors, identify and correct type definitions for type errors, follow the debugging skill for test failures, and check configuration and dependencies for build errors. After the fixes are made, it automatically commits them and retriggers CI, completing the automated repair loop.