Automate CI/CD and Development Workflows

Workflow Automation is a skill for designing and building CI/CD pipelines, GitHub Actions workflows and other automated development processes that cut manual work, improve consistency and speed up delivery while keeping quality and security gates in place. Its instructions call for inventorying current build, test and deploy steps and target environments, then defining pipeline stages with caching, artifacts and quality gates, adding security scans, secret handling and approvals for risky steps, and documenting a rollout, rollback and notification strategy. Its required output format is a summary of pipeline stages and triggers, proposed workflow files or a step list, the required secrets, environment variables and service integrations, and explicit risks, assumptions and rollback notes.

The bundled resources/implementation-playbook.md fills this out with concrete patterns:

• A workflow-discovery approach that scans a project for existing GitHub Actions, GitLab CI or Jenkins pipelines, flags manual build/deploy/release/test shell scripts as high automation potential, and checks README files for documented manual steps, then generates prioritized recommendations (for example: no CI/CD found is flagged high priority for GitHub Actions, GitLab CI or Jenkins; manual deployment is flagged critical priority for tools like ArgoCD, Flux or Terraform).
• A multi-environment GitHub Actions pipeline pattern with separate quality, test, build, deploy and post-deployment-verify jobs: quality runs linting, type checking, a security audit and a license check; test runs a matrix across operating systems and Node versions with unit and integration tests and coverage upload; build produces per-environment Docker images, scans them for vulnerabilities, and pushes to a registry; deploy updates an ECS task definition and service per environment and posts a Slack notification; verify runs smoke tests, Cypress end-to-end tests, a sitespeed.io performance budget check and a ZAP baseline security scan against the deployed URL.
• A semantic-release workflow that runs npx semantic-release on pushes to main, gated by a .releaserc.js config naming branches (including beta and alpha prereleases) and plugins for commit analysis, release notes, a changelog file, npm publish, a git commit of the changelog and package.json, and a GitHub release.
• A pre-commit hook configuration chaining hooks for trailing whitespace, large-file and merge-conflict checks, private-key detection, Python formatting with Black and isort, flake8 linting, ESLint and Prettier for JS/TS, and a local unit-test hook that runs on commit.

It extends further into infrastructure, monitoring, dependency, documentation and security automation. A Terraform workflow runs terraform fmt -check, init, validate and plan, posts the plan summary as a PR comment, and applies on merge to main. A monitoring workflow installs the kube-prometheus-stack with Helm, applies Grafana dashboards and alert rules with kubectl apply from repo paths, and sets up Alertmanager routing with Helm. A Renovate configuration extends the automerge-digest and automerge-minor presets, auto-merges devDependencies and @types/* packages, groups ESLint packages, pins Docker digests, auto-merges security vulnerability alerts, and runs on weekday nights (after 10pm and before 5am) and weekends. A documentation workflow generates TypeDoc API docs and Mermaid architecture diagrams and deploys the built site to GitHub Pages. A security-scanning workflow chains Trivy (filesystem scan, SARIF upload), Snyk, OWASP Dependency-Check, SonarCloud, Semgrep (security-audit/secrets/OWASP-Top-Ten rulesets) and GitLeaks secret scanning, run on push, PR and a weekly Sunday cron. A TypeScript WorkflowOrchestrator class models parallel or sequential steps with retries, timeouts, conditions and a per-step error policy of fail, continue or retry.

Use it for automating CI/CD workflows or release pipelines, designing GitHub Actions or multi-stage build/test/deploy flows, replacing manual build, test or deployment steps, or improving pipeline reliability, visibility or compliance checks. Do not use it for a one-off command or quick troubleshooting, when there is no workflow or automation context at all, or for a task that is strictly product or UI design. Its safety note warns against running deployment steps without approvals and rollback plans, and to treat secrets and environment configuration changes as high risk.

Input is the project's current build, test and deploy setup plus target environments (the $ARGUMENTS the skill is invoked with). Output is the four-part summary above: pipeline stages and triggers, proposed workflow files or steps, the secrets/env vars/integrations they need, and risks, assumptions and rollback notes.

Its examples integrate with GitHub Actions, GitLab CI and Jenkins for pipeline execution; Codecov for coverage reporting; Trivy and OWASP ZAP for container and web security scanning; AWS ECS for deployment; Slack for deployment notifications; Cypress and sitespeed.io for end-to-end and performance verification; and semantic-release with npm and GitHub release plugins for versioned releases. Pre-commit hooks integrate Black, isort, flake8, ESLint and Prettier.

It suits developers and platform engineers who need to design or harden CI/CD pipelines, replace manual release steps with automation, or add security scanning, approvals and rollback plans to an existing delivery process.