ci-cd-pipeline_skill

Overview

This skill documents CI/CD Pipeline Management for a TypeScript-backed project using GitLab CI/CD. It focuses on pipeline architecture, build-time bottlenecks, caching strategies, and a practical path to a 3-stage deployment workflow that reduces build time by up to 60–65%. The goal is faster, more reliable builds with minimal ongoing maintenance.

How this skill works

The skill inspects the GitLab CI configuration, the build stage steps, and the before_script actions to identify costly operations like compiling PHP extensions every run. It recommends a three-layer caching strategy (APT, Node modules, Composer vendor) and a custom Docker image that ships pre-compiled extensions and system libraries. It also describes how to update .gitlab-ci.yml and push a one-time image to the GitLab Container Registry.

When to use it

• When you need to reduce CI build time and cost
• When builds fail due to missing system libraries or extension compile errors
• When you want deterministic, repeatable build environments
• When planning a staged rollout: validate → build → deploy
• When adding caching to speed dependency installs and asset builds

Best practices

• Build and publish a custom CI image with pre-installed system libs and PHP extensions
• Use cache keys tied to lockfiles (package-lock.json, composer.lock) to avoid stale caches
• Keep git submodule init in before_script for VCS path packages
• Test CI changes in a branch and monitor cache hit rates and stage durations
• Limit one-time setup in pipeline jobs and shift repeated work into the image

Example use cases

• Cutting build time from ~12 minutes to ~3–4 minutes by using a custom Docker image
• Fixing docker-php-ext-install failures by adding missing apt packages to the image
• Recovering from npm ci failures by regenerating package-lock.json and committing it
• Triggering staging deploys manually and production deploys via Git tags
• Tracking pipeline health via average build time, cache hit rate, and deploy frequency