How To Use Spec-Driven Development for Production Workflows
summary
Spec-driven development breaks complex features into requirements, technical design, and implementation steps, using specs as the primary contract so AI agents stay on target. The workflow demo uses the Kiro agentic IDE (and CLI) with an MCP hookup, agents.md steering files and skills, builds an '80s‑themed movie database that starts with broken links (404s) then iterates to a working MVP, and validates behavior with property‑based tests using fast-check while keeping a human reviewer in the loop.
AI-written summary · about
chapters (10)
- 0:00What Is Spec-Driven Development?
- 1:59Why AI Agents Need Better Requirements
- 3:54Meet Eric and the Case for Planning First
- 4:53Can't Frontier Models Do Everything?
- 6:00AI Coding Principles: Context, Skills, and Trust
- 9:25Speed vs. Maintainability
- 10:56How Kiro and Spec-Driven Development Evolved
- 13:08The Spec-Driven Development Workflow
- 17:59MCP, Product Specs, and Team Collaboration
- 21:21Live Demo: Building an App with Spec Mode
related talks
18mUsing Spec-Driven Development for Production Workflows
47mSpec-Driven Development with AI Agents From High-Level Requirements to Working SW
25mspec-driven development
45mSpec-Driven Development in the Age of Agents
01h 04mSpec-Driven Development: Agentic Coding at FAANG Scale and Quality
45mSpec Driven Development with Kiro: Ensuring Quality and Traceability in the AI Era
50mSpec-driven Development using Coding Agents
37mSpec Driven Development: AI Assisted Coding With Doc at the Core