Kiro New2025年〜
An AI-powered IDE that enables spec-driven development
What It Does
Kiro is an AI-powered integrated development environment (IDE) from AWS, built around spec-driven development. AI supports the entire development lifecycle - from requirements definition, design, task management, code generation, to test creation. It automatically generates design documents and task lists from requirements written in natural language, and AI agents execute each task. Agent hooks enable automated actions on file save or commit, and steering files let you define project-specific rules for the AI.
Use Cases
Used for end-to-end development from requirements definition to implementation for new projects, refactoring and bug fixing in existing codebases, automated test code generation, automated code reviews, automated documentation generation, quality assurance through property-based testing, and creating and managing infrastructure as code (IaC).
Everyday Analogy
Think of it like an excellent pair programmer. Not only do they sit beside you and write code together, but they also organize project requirements, draft design documents, break down tasks, and even write tests. Plus, they remember your project rules (steering) and automatically check (hook) every time you save a file.
What Is Kiro?
Kiro is an AI-powered IDE announced by AWS in 2025. Based on a VS Code-compatible editor, it provides a unique workflow called spec-driven development. When a developer describes requirements in natural language, the AI automatically generates a requirements document (requirements.md), design document (design.md), and task list (tasks.md), then executes each task in order to generate code. It's not just a code completion tool - it's a development environment that supports the entire software development lifecycle.
Spec-Driven Development
Spec-driven development, Kiro's core feature, provides AI-assisted structured workflow through requirements definition, design, task breakdown, and implementation. Artifacts from each phase are saved as markdown files and can be version-controlled. Design documents also include property-based test specifications for formally verifying implementation correctness. This approach significantly improves the quality and consistency of AI-generated code.
Agent Hooks and Steering
Agent hooks are a feature that triggers AI agent actions on specific events (file save, file creation, commit, etc.). For example, you can automatically run linting every time a TypeScript file is saved, or automatically run tests when a test file is created. Steering files are configuration files that communicate project-specific rules and coding conventions to the AI, and can be managed at both global and workspace levels. For case studies and usage tips on agent hooks and steering, specialized books (Amazon) can also be a reference.
Getting Started
To get started with Kiro, download and install it from the official website. Existing VS Code extensions work as-is, and settings and keybindings carry over. Once you open a project, simply describe your requirements in natural language from the chat panel to start spec-driven development. Configure steering files to define project rules and set up agent hooks for automation to further boost development efficiency.
Things to Watch Out For
- Being VS Code-compatible, existing extensions and keybindings work as-is, though some extensions may have compatibility issues
- The quality of code generated through spec-driven development heavily depends on steering file configuration, so it's important to carefully define rules suited to your project
- AI agent execution consumes credits, so estimate consumption in advance for large-scale tasks
