liveBook · Manning

Specification-driven development

As explained in Context Engineering

Spec-driven development is a software development approach in which a specification guides planning, task breakdown, and implementation.

Specification-driven development (SDD) organizes AI-assisted development around a specification rather than a succession of disposable prompts. The spec supplies a stable context source for a coding agent, capturing intent, constraints, interfaces, edge cases, validation criteria, and nonfunctional expectations before implementation begins. section 11.3.7

A prompt-first loop can be fast for small tasks, but becomes harder to manage across multiple files, roles, or later maintenance. A specification narrows ambiguity and gives the agent a shared decision-making frame for planning, coding, testing, review, and change. section 11.3.7

The relationship between the specification and the codebase can take three forms:

  • Spec-first: Write the specification before engineering work and use it to guide an implementation task. After generation, engineers typically discard the spec and maintain the code. section 11.3.7
  • Spec-anchored: Keep the specification aligned with future changes. section 11.3.7
  • Spec-as-source: Treat the specification as the primary artifact from which implementation is generated. section 11.3.7

Backlog-scoring feature

A staged workflow separates three reviewable artifacts: a specification describing what the feature must do and why, a plan recording the technical approach, and a task list turning the plan into executable work. The agent resolves ambiguity, checks consistency, and implements after the checks are complete. section 11.5.1

Listing 11.2 Compact Spec Kit workflow for the backlog-scoring example
specify init . #1
/speckit.constitution #2
/speckit.specify #3
/speckit.clarify #4
/speckit.plan #5
/speckit.tasks #6
/speckit.analyze #7
/speckit.implement #8

The listing shows the Spec Kit initialization and slash-command sequence for the staged workflow. section 11.5.1

Run the slash commands from a supported coding agent after initializing Spec Kit in the repository. Keeping the artifacts inspectable and reusable makes the work easier to review than relying on a long conversational transcript. section 11.5.1

How Context Engineering covers it

How deep it goes
Distinguishes three specification–code relationships and walks through a staged workflow from captured intent to implementation.
Examples
Uses an API endpoint specification and a backlog-scoring feature to illustrate constraints, plans, and executable tasks.
What you can do after
Apply it via GitHub Spec Kit using staged commands for specification, clarification, planning, tasks, analysis, and implementation.
How it connects
Connects specification-driven development with context engineering, coding agents, and the software development lifecycle.

Also explained in

Compare how 4 books explain specification-driven development →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Context Engineering. The section links point to the parts of the book it draws on.