Skip to main content

Chapter 31: Spec-Kit Plus Hands-On

Welcome to hands-on specification-driven development. This chapter transforms understanding (from Chapter 30) into practice. You won't read about specifications-you'll write them. You won't learn about AI collaboration - you'll build project with your AI companion from specifications to implementation.

From Chapter 30 to Chapter 31​

Chapter 30 taught you the foundations:

  • Why specification-driven development matters (the vague code problem)
  • What SDD is (spec-first, spec-anchored, spec-as-source levels)
  • How to write your first spec (calculator exercise)
  • Why teams need shared rules (Constitutions, ADRs, PHRs)
  • When SDD became essential (AI convergence moment)
  • Which tools exist (Kiro, Spec-Kit, Spec-Kit Plus, Tessel)

Chapter 31 teaches you the workflow:

  • How to install and configure Spec-Kit Plus
  • How to create project Constitutions
  • How to collaborate with AI on specifications
  • How to refine specs iteratively
  • How to generate plans and document decisions (ADRs)
  • How to decompose work with checkpoint-driven execution
  • How to implement with validation protocols
  • How to complete end-to-end workflows independently

The Bridge: Chapter 30 gave you the map. Chapter 31 teaches you to drive.

What You'll Learn​

By completing this chapter, you'll master:

  1. Spec-Kit Plus Architecture - Understand Horizontal Intelligence (ADRs + PHRs) and Vertical Intelligence (orchestrator + subagents)
  2. Evals-First Collaboration - How to explore problem space with AI BEFORE writing formal specifications
  3. Constitution Creation - Setting project-wide quality standards that cascade to all features
  4. Specification Writing - Creating clear, testable requirements using SMART criteria
  5. Iterative Refinement - Using /sp.clarify to identify and resolve specification gaps
  6. Architecture Planning - Generating implementation plans and documenting decisions with ADRs
  7. Task Decomposition - Breaking plans into atomic work units with checkpoint-driven execution
  8. AI-Driven Implementation - Orchestrating code generation with systematic validation protocols
  9. Independent Workflow - Applying all skills end-to-end without guidance

Most importantly, you'll build the muscle memory for specification-first thinking that makes you dramatically faster in Chapters 32 and beyond. Let's build something real.