Skip to content

Chapter 7 - Large Feature Development

Implement multi-file changes using AI assistance to plan, execute, and test complex features across a full-stack application.

  • Reading time: 45 minutes
  • Practical time: 6-8 hours (full day)

Reading

Briefs

Assets

Feature stories and requirements:

⚠️ Important: These assets contain feature stories and requirements. No configuration files need to be overwritten for this chapter.

Practical

Task: Implement a Large Feature

Choose one story from the list below. Each story includes functional requirements, technical requirements, and acceptance criteria.

Choosing a Story

All stories are available in the Assets section above.

StoryComplexityStack
UI Internationalization (i18n)ModerateFrontend
Accessibility MenuModerateFrontend
Image Upload SupportHighFull-stack
Order Status WorkflowHighFull-stack
Real-Time Product UpdatesHighFull-stack
Docker ImagesModerateDevOps
Keycloak OAuth AuthenticationHighFull-stack
Multi-Location Fulfillment StrategyModerateBackend
NgRx SignalStore MigrationHighFrontend

Recommended for first attempt: UI Internationalization - moderate complexity

Workflow

For your chosen story, follow these steps:

  1. Read the story: Understand the functional and technical requirements, scope, and acceptance criteria
  2. Explore options: Discuss implementation approaches with the agent (pros, cons, trade-offs for a small project)
  3. Research libraries: Identify production-ready options for the technical decisions in your story
  4. Create a plan: Break the work into phases with clear milestones
  5. Implement incrementally: Complete one phase at a time. Commit after each phase.
  6. Test thoroughly: Run manual tests and automated tests. Verify each acceptance criterion is met

What to Observe

  • How agents help break down requirements into implementation steps
  • Planning and architectural decision support quality
  • Managing multi-file changes with version control
  • When you need to steer the agent vs. when it makes good decisions independently

Continue with: Spec-Driven Agentic Coding with OpenSpec — learn how to use structured change proposals (/opsx:propose → review → /opsx:apply → /opsx:archive) to drive AI agents through larger features with more consistency.

Copyright © MSG Systems Romania AI Labs