You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have massivly used this approach in the past to first craft an initital feature promt, where LLM is supporting me with questions. Sometimes even the initial idea has some weaknesses and AI can help to craft out the idea. I created a more bad then good /ideate-feature command for claude which I did not spent enough brain into, to say its a good one, but it works. It was relying on this approach https://github.com/coleam00/context-engineering-intro/tree/main. Does it make sense to have something similar for spec-kit? Like a / command which helps with feature ideation?
Currentl /ideate-feature
# Ideate Feature
I'll help you define a new feature for the platform through an iterative, conversational process. We'll develop a comprehensive feature specification that can be implemented following our established patterns and architecture.
<context>
This is a Next.js 15 application for that manages:
- <ADD SOMETHING HERE>
Tech Stack: Next.js 15, React 18, TypeScript, MUI v7, Prisma/PostgreSQL, NextAuth.js, Vercel Blob Storage
</context>
<process>
1. I'll ask you one short, precise and single sentance question at a time to understand your feature idea. Questions should be asked to give short replies.
2. Each question builds on your previous answers
3. I'll incorporate relevant project context and constraints
4. If existing feature is referenced ask to add the corresponding file into context. Ask about source file to be added to the context.
5. We'll iteratively refine the feature until we have a complete specification
6. The final output will be a feature document in PRPs/feature/[feature-name].md using the "Feature Definition Template"
</process>
<ideation-approach>
I'll guide you through:
- Understanding the core problem/need
- Identifying target users and their workflows
- Exploring integration with existing features
- Considering technical constraints and patterns
- Planning data models and API requirements
- Designing user interfaces and interactions
- Addressing security and permission requirements
- Defining success metrics and validation
</ideation-approach>
<thinking>
Before asking my first question, I need to understand:
- What problem this feature solves
- Who will use it
- How it fits into the existing platform
- What similar patterns we can leverage
</thinking>
## Feature Definition Template
Once we've gathered all requirements through our conversation, I'll create a feature specification using this structure:
## FEATURE: [Feature Name]
[Comprehensive description of the feature, its purpose, and value proposition]
### USER STORIES
[Detailed user stories covering different personas and use cases]
### INTEGRATION POINTS
[How this feature connects with existing functionality:
- Appointments system
- Groups and status reports
- Newsletter system
- Admin workflows
- Authentication and permissions]
### DATA MODEL
[Required database models, relationships, and fields]
### API REQUIREMENTS
[API endpoints, request/response formats, validation rules]
### UI/UX DESIGN
[Interface requirements, component patterns, user flows]
### SECURITY & PERMISSIONS
[Access control, data privacy considerations, authorization rules]
### EXAMPLES
[Concrete examples of the feature in action, including:
- Sample data
- User workflows
- Edge cases]
### DOCUMENTATION
[Resources and references needed:
- Similar features in other systems
- Technical documentation
- Design patterns to follow]
### OTHER CONSIDERATIONS
[Technical constraints, performance requirements, future extensibility]
<instructions>
Initial feature idea: $ARGUMENTS
Let's begin our ideation process. Remember, I'll ask one question at a time to help us thoroughly explore and define your feature idea.
</instructions>
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I have massivly used this approach in the past to first craft an initital feature promt, where LLM is supporting me with questions. Sometimes even the initial idea has some weaknesses and AI can help to craft out the idea. I created a more bad then good /ideate-feature command for claude which I did not spent enough brain into, to say its a good one, but it works. It was relying on this approach https://github.com/coleam00/context-engineering-intro/tree/main. Does it make sense to have something similar for spec-kit? Like a / command which helps with feature ideation?
Currentl /ideate-feature
Beta Was this translation helpful? Give feedback.
All reactions