Configure Bitbucket CI/CD Pipelines

This skill turns the assistant into an expert for building efficient, scalable Bitbucket Pipelines configurations in bitbucket-pipelines.yml. It covers core pipeline structure (default and branch-specific pipeline steps), multi-stage pipelines with parallel step execution, advanced caching strategies (custom cache definitions for tools like Gradle, Maven, Cypress, and Next.js, plus cache-optimization patterns using --prefer-offline and pip's cache directory), Docker integration (service containers such as Postgres for integration tests, building and pushing images, and multi-stage Docker builds with build-args), deployment patterns to AWS (S3 sync plus CloudFront invalidation, with manual-trigger production deploys) and Kubernetes (fetching kubectl, decoding a base64 kubeconfig secret, templating an image tag, and checking rollout status), and a set of best practices for secrets management, performance, error handling, and conditional execution based on changed paths.

Use it when writing or restructuring a bitbucket-pipelines.yml file and you need concrete, working YAML: a basic build-test-deploy pipeline, a parallel unit-test-plus-lint-and-security-scan stage feeding into a build step, custom cache definitions, a Postgres service container for integration tests, single- and multi-stage Docker build-and-push scripts using $BITBUCKET_REPO_FULL_NAME and $BITBUCKET_COMMIT, AWS S3/CloudFront and Kubernetes deployment scripts, environment-variable validation guards, step size limits for memory-intensive tasks, and condition.changesets.includePaths for path-based conditional execution. It is not a general CI/CD theory guide and does not cover other CI platforms (GitHub Actions, GitLab CI, Jenkins) - the YAML and environment variables shown are Bitbucket Pipelines-specific.

Inputs are the project's build/test/deploy commands and its target deployment infrastructure (Docker registry, AWS account, or Kubernetes cluster). Outputs include ready-to-adapt pipeline YAML for each covered scenario: a parallel block running independent Unit Tests and Lint/Security Scan steps before a Build Application step with an after-script listing the build output, a definitions.caches block for custom cache paths, a definitions.services.postgres block with environment variables, Docker build/tag/push scripts including a multi-stage build with --target production and --build-arg, AWS CLI deployment commands, and a Kubernetes deploy script that downloads kubectl, writes the kubeconfig, substitutes the image tag with sed, and runs kubectl rollout status.

image: node:18

pipelines:
  default:
    - step:
        name: Build and Test
        caches:
          - node
        script:
          - npm ci
          - npm run test
          - npm run build
        artifacts:
          - dist/**
  branches:
    master:
      - step:
          name: Deploy to Production
          deployment: production
          script:
            - echo "Deploying to production"
    develop:
      - step:
          name: Deploy to Staging
          deployment: staging
          script:
            - echo "Deploying to staging"

The patterns target Bitbucket Pipelines' own step, cache, service, and deployment primitives, integrating with Docker (build/push to a configurable registry), AWS (S3 and CloudFront via the AWS CLI), and Kubernetes (via kubectl against a base64-encoded kubeconfig secret). A step's size: 2x setting doubles available memory to 4GB (8GB when a Docker service is running), letting a memory-heavy caching or build step opt in without changing the pipeline's global defaults.

Teams setting up or optimizing a Bitbucket Pipelines CI/CD workflow, especially those adding parallel test execution to cut build time, wiring a Docker build-and-push step, or configuring environment-specific (staging vs. production, manual-trigger) deployments to AWS or Kubernetes. Also useful for anyone hardening an existing pipeline around secrets handling, memory-intensive step sizing, and path-based conditional runs.