partme-ai/gitlab-ci

When to use this skill

Use this skill whenever the user wants to:

• Write or debug GitLab CI pipelines (.gitlab-ci.yml)
• Configure GitLab Runner, stages, jobs, and artifacts
• Set up CI/CD variables, secrets, and environments
• Integrate testing, building, and deployment automation
• Use needs for complex dependency graphs between jobs

How to use this skill

Workflow

1. Define stages — declare the pipeline execution order
2. Write jobs — assign each job to a stage with scripts and rules
3. Configure artifacts and cache — pass build outputs between jobs
4. Set up environments — define deployment targets with rules

Quick Start Example

# .gitlab-ci.yml
stages:
  - test
  - build
  - deploy

variables:
  NODE_VERSION: "20"

test:
  stage: test
  image: node:${NODE_VERSION}
  cache:
    key: ${CI_COMMIT_REF_SLUG}
    paths:
      - node_modules/
  script:
    - npm ci
    - npm test
  artifacts:
    when: on_failure
    paths:
      - test-reports/
    expire_in: 7 days

build:
  stage: build
  image: node:${NODE_VERSION}
  needs: [test]
  script:
    - npm ci
    - npm run build
  artifacts:
    paths:
      - dist/

deploy_production:
  stage: deploy
  needs: [build]
  environment:
    name: production
    url: https://myapp.example.com
  script:
    - ./scripts/deploy.sh
  rules:
    - if: $CI_COMMIT_BRANCH == "main"

Key Concepts

| Feature | Purpose | |---------|---------| | stages | Define execution order | | needs | Create DAG dependencies (skip stage waiting) | | artifacts | Pass files between jobs | | cache | Speed up repeated installs | | rules | Control when jobs run | | environment | Track deployment targets |

Best Practices

• Use stages to define clear build order; jobs within a stage run in parallel
• Store sensitive values in CI/CD Variables (Settings > CI/CD > Variables) — never hardcode in YAML
• Use needs to create complex dependency graphs and avoid unnecessary waiting
• Preserve logs and artifacts on failure with artifacts: when: on_failure
• Use rules instead of deprecated only/except for conditional job execution

Troubleshooting

• Job stuck pending: Check runner tags match and runners are available
• Artifact not found: Verify the producing job completed and artifact paths are correct
• Cache not restoring: Ensure cache key is consistent; check runner cache configuration
• Pipeline not triggered: Verify rules conditions match the event (push, merge request, etc.)

Keywords

gitlab ci, gitlab-ci, pipeline, runner, ci/cd, artifacts, cache, environments, deployment automation