The Plan agent is a specialized built-in agent that helps transform ideas into structured implementation plans. It guides you through requirements gathering, research, and creates detailed task breakdowns before switching to execution mode.
Press Shift + Tab to toggle between plan and execution modes.
> /plan Switched to the Kiro [plan] agent. Transform any idea into fully working code. What do you want to build today? [plan] >
> /plan Build a REST API for user authentication
When active, you'll see:
[plan] indicator in your promptThe planner guides you through structured questions to refine your initial idea.
[plan] > I want to build a todo app I understand you want to build a todo app. Let me help you plan this implementation. **My Understanding:** You want to create a todo application for managing tasks. Let me gather some requirements to create a detailed plan: [1]: What platform should this todo app target? a. **Web Application** - Browser-based with HTML/CSS/JavaScript b. **Mobile App** - Native iOS/Android or cross-platform c. **Desktop App** - Electron, Tauri, or native desktop d. **CLI Tool** - Command-line interface e. **Other** - Provide your own answer [2]: What's the primary use case for this todo app? a. **Personal Task Management** - Individual productivity b. **Team Collaboration** - Shared tasks and assignments c. **Project Management** - Complex workflows with dependencies d. **Other** - Describe your specific needs (Use the chat to answer any subset: eg., "1=a, 2=b" or provide your own answers)
| Feature | Description |
|---|---|
| Structured questions | Clear, numbered questions with multiple choice options |
| Flexible responses | Answer individual questions or provide custom responses |
| Iterative process | Builds understanding through multiple rounds of questions |
| Context awareness | Adapts follow-up questions based on your answers |
The planner explores your codebase and researches relevant technologies.
| Capability | Description |
|---|---|
| Code exploration | Uses code, grep, and glob tools to understand existing code |
| Technology research | Identifies relevant frameworks, libraries, and patterns |
| Architecture analysis | Reviews existing project structure and conventions |
Creates a detailed, step-by-step implementation plan with clear objectives.
**Implementation Plan - Todo CLI Command** **Problem Statement:** Add todo management to existing Kiro CLI for task tracking. **Requirements:** - CLI commands for CRUD operations - Local SQLite storage - Priority and due date support **Task Breakdown:** Task 1: Create database schema and models - Define Todo struct with required fields - Create database migration for todos table - Demo: Can create and query todos in database Task 2: Implement CLI command structure - Add todo subcommand with add/list/complete operations - Demo: CLI accepts todo commands and shows help Task 3: Add advanced features - Implement due dates and priority sorting - Demo: Complete todo system with all features
Each task includes:
Before switching to execution, the planner seeks your approval.
[plan] > Does this plan look good, or would you like me to adjust anything? > The plan looks great! Let's implement it. *Planning complete!* Ready to exit [plan] agent to start your implementation? [y/n]: > y
The handoff process:
The Plan agent operates in read-only mode to keep focus on planning.
| Operation | Status |
|---|---|
| File reading | ✓ Full access to explore codebase |
| Code intelligence | ✓ LSP operations for understanding code structure |
| Search (grep, glob) | ✓ Code exploration |
| Web search | ✓ Documentation and research access |
| File writing | ✗ Cannot create or modify files |
| Command execution | ✗ Limited to read-only bash commands |
| MCP tools | ✗ Cannot use MCP tools |
| Issue | Solution |
|---|---|
| How to exit plan agent? | Use Shift + Tab to return to your previous agent. |
| Planning session interrupted | Use /plan to re-enter planning and continue where you left off. |
| Plan not transferring to execution agent | Make sure to respond "y" when prompted with "Ready to exit [plan] agent to start your implementation? [y/n]". |
> /plan Add user authentication to my web app [plan] > I understand you want to add user authentication to your web application. **My Understanding:** You want to implement user authentication functionality for an existing web application. [1]: What authentication method do you prefer? a. **Email/Password** - Traditional username and password b. **OAuth** - Google, GitHub, or other OAuth providers c. **Magic Links** - Passwordless email-based authentication d. **Multi-factor** - Combination of methods with 2FA > 1=a [plan] > Great! Email/password authentication it is. [2]: What's your current tech stack? a. **React + Node.js** - JavaScript full-stack b. **Next.js** - React with server-side rendering c. **Django/Flask** - Python web framework d. **Other** - Please specify > 2=d, I'm using Rust with Axum framework [plan] > Perfect! Rust with Axum is an excellent choice for secure authentication. *Researching Axum authentication patterns...* *Exploring your existing codebase structure...* **Implementation Plan - User Authentication System** [Detailed plan follows...] Does this plan look good, or would you like me to adjust anything? > Looks perfect! Ready to exit [plan] agent to start your implementation? [y/n]: y [default] > Implement this plan: [Plan content transferred to execution agent]
Plan agent