Orchestrate Full Stack Feature Development

This skill orchestrates full-stack feature delivery across four phases by dispatching work to specialized subagents in sequence and in parallel. Phase 1 (Architecture & Design) runs a database-architect to design schema and migrations, a backend-architect to define service boundaries and OpenAPI/GraphQL contracts using that schema, and a frontend-developer to design component hierarchy and state management against those contracts. Phase 2 (Parallel Implementation) runs backend implementation (python-pro, golang-pro, or nodejs-expert depending on stack), frontend implementation (React/Next.js components with Storybook stories and WCAG 2.1 AA accessibility), and database implementation/optimization (sql-pro) concurrently, since all three consume the same Phase 1 architecture but don't depend on each other's output.

Use it for delivering a complete feature spanning backend, frontend, and infrastructure with API-first, contract-driven development - not for a single-layer change that doesn't need cross-layer coordination. It assumes a multi-subagent environment where each named subagent_type (database-architect, backend-architect, test-automator, security-auditor, deployment-engineer, performance-engineer, etc.) is actually available to dispatch to. Coordination notes call out that Phase 2's three implementation tasks can run simultaneously but must all converge before Phase 3 begins, and that consistent error handling and API response shapes should be maintained across every service the phases touch.

The feature description is passed as $ARGUMENTS into each phase's prompt, and each step's expected output feeds the next: Phase 3 (Integration & Testing) runs API contract tests with Pact/Dredd, end-to-end tests with Playwright/Cypress covering critical journeys and visual regression, and a security audit against OWASP Top 10 plus XSS/CSRF and secrets-management review. Phase 4 (Deployment & Operations) sets up Docker/Kubernetes and CI/CD with feature flags (LaunchDarkly/Unleash) and blue-green deployment, wires observability (OpenTelemetry tracing, Prometheus/DataDog metrics, ELK/Splunk logging, SLIs/SLOs), and optimizes performance (caching, CDN, bundle size) using data from the observability step. Configuration options include stack, deployment_target, feature_flags, api_style (REST or GraphQL), testing_depth, and compliance (GDPR/HIPAA/SOC2).

Coordinates named subagent types across database design, backend and frontend development, testing, security, and deployment; the tooling it names spans OpenAPI/GraphQL, Redux/Zustand/Context, Pact/Dredd, Playwright/Cypress, LaunchDarkly/Unleash, OpenTelemetry, Prometheus/DataDog, ELK/Splunk, and Terraform/CloudFormation for infrastructure-as-code.

A team or orchestrator building a complete feature who needs a coordinated hand-off between database, backend, and frontend design through parallel implementation to testing and production deployment - tracking success against ten explicit criteria (contract tests validated, integration and E2E tests passing, a clean security audit, SLOs met, observability capturing key metrics, feature flags for progressive rollout, complete documentation, automated CI/CD quality gates, and verified zero-downtime deployment) and maintaining traceability via correlation IDs across all services and architectural decisions recorded as ADRs throughout the whole delivery, from the initial database schema decision to the final rollout configuration.