Automation tool for peripheral tasks in AI-driven development
Team Kit is a collection of commands that works with AI editors like Claude to automate everything from requirements definition to specification creation and mockup generation.
- Automated generation of requirements definitions
- Extraction of use cases
- Generation of UI definitions
- Creation of screen flow diagrams
- Automated generation of HTML mockups
- Collection and reflection of user feedback
- Maintenance of consistency between requirements and mockups
- Claude Code
- bash shell environment
# Install to current directory
curl -fsSL https://raw.githubusercontent.com/tango238/teamkit/main/install.sh | bash -s -- .# Force overwrite (no confirmation)
curl -fsSL https://raw.githubusercontent.com/tango238/teamkit/main/install.sh | bash -s -- --yes .
# Install to a specific directory
curl -fsSL https://raw.githubusercontent.com/tango238/teamkit/main/install.sh | bash -s -- /path/to/project
Options:
--yes,-y,--force,-f: Overwrite existing files without confirmation
The installation script copies all command files under the .claude/commands/teamkit directory to the specified project directory with the same structure.
Team Kit provides a step-by-step specification creation workflow. Each step is available as a slash command /teamkit:*.
First, create a directory to manage specifications:
your-project/
└── .teamkit/
└── YourFeature/
└── README.md # Describe requirements
README.md example:
# Facility Management
## Background
Management of basic information for facilities, room types, and rooms is essential for facility operations. This information forms the basis for reservation management and pricing, so it is necessary to keep it accurate and up-to-date.
## Objectives
- Centralize the management of basic facility information to enable efficient operation of multiple facilities
- Manage the characteristics of each room type to meet diverse accommodation needs
- Appropriately manage the status of each room to accurately determine availability and cleaning requirements
## Main Actors
- Facility Manager
- Cleaning Staff
## Business Overview
Facility management involves the registration, update, deletion, and retrieval of facilities, room types, and rooms. Facility information includes name, address, check-in/check-out times, etc., while room types include capacity and smoking policy. Each room is linked to a room number and room type, and availability is controlled by its status.
## Requirements
### Facility Manager
Register and update facility information, room types, and room information
Extract requirements from README.md and generate workflow definitions:
/teamkit:generate-workflow YourFeature
Generated Files:
.teamkit/YourFeature/workflow.yml- Workflow definition.teamkit/YourFeature/status.json- Status management file
Generate all specifications and HTML mockups from the README in one command:
/teamkit:generate YourFeature
Generated Files:
.teamkit/YourFeature/workflow.yml- Workflow definition.teamkit/YourFeature/usecase.yml- Use cases.teamkit/YourFeature/ui.yml- UI definition.teamkit/YourFeature/screenflow.md- Screen flow diagram.teamkit/YourFeature/mock/*.html- Mockups for each screen
Options:
/teamkit:generate YourFeature --manual # Also generate user manual
/teamkit:generate YourFeature --test # Also generate acceptance tests
/teamkit:generate YourFeature --all # Generate manual + acceptance tests
Additional generated files (with options):
.teamkit/YourFeature/manual.md- User operation manual (--manualor--all).teamkit/YourFeature/acceptance-test.md- Acceptance test cases (--testor--all)
If you already have ui.yml and want to regenerate mockups:
/teamkit:create-mock YourFeature
Generate manual and acceptance test items along with the standard pipeline:
/teamkit:generate YourFeature --all
Submit feedback on specifications:
/teamkit:feedback YourFeature --preview "Please split the address field into details"
When you submit feedback, the AI checks the scope of impact and creates details and TODOs in feedback.md.
Review the content and mark TODOs with [o] if you want to apply them.
If you add the -p or --preview option, the changes will be immediately reflected in the mockup, allowing you to confirm the feedback content.
You can adjust the feedback application content by changing the Next action in the Summary.
Apply feedback:
/teamkit:apply-feedback YourFeature
Check current step information:
/teamkit:get-step-info YourFeature
After installation, the project will have the following structure:
your-project/
├── .claude/
│ └── commands/
│ └── teamkit/ # Team Kit commands
│ ├── generate-workflow.md
│ ├── generate-usecase.md
│ ├── generate-ui.md
│ ├── generate-screenflow.md
│ ├── generate-mock.md
│ ├── generate-manual.md
│ ├── generate-acceptance-test.md
│ ├── generate.md
│ ├── create-mock.md
│ ├── feedback.md
│ ├── apply-feedback.md
│ ├── get-step-info.md
│ └── update-status.md
└── .teamkit/
└── <feature-name>/
├── README.md # Requirements definition
├── workflow.yml # Workflow definition
├── usecase.yml # Use cases
├── ui.yml # UI definition
├── screenflow.md # Screen flow diagram
├── status.json # Status management
├── feedback.md # Feedback
├── manual.md # User operation manual (optional)
├── acceptance-test.md # Acceptance test cases (optional)
├── mock/screens.yml # Screen generation status
└── mock/*.html # Mockups for each screen
Example of a typical development flow:
# 1. Describe requirements in README.md
# 2. Generate all specifications and mockups
/teamkit:generate OrderManagement
# 3. Or generate with manual and acceptance tests
/teamkit:generate OrderManagement --all
# 4. Check the generated mockup
# Open .teamkit/OrderManagement/index.html in a browser
# 5. Submit feedback if any and check the mockup
/teamkit:feedback OrderManagement -p "Please add an order cancellation function"
# 6. Apply feedback
/teamkit:apply-feedback OrderManagement
# 7. Regenerate mockups only (after manual ui.yml edits)
/teamkit:create-mock OrderManagement- Command Descriptions: English
- Generated Specifications: Japanese
- Status Messages: Japanese
This allows the LLM to understand accurately while generating specifications in Japanese.
See the LICENSE file for the license of this project.
If you encounter any issues or have feature requests, please report them in the Issues section on GitHub.