| name | feature-design-assistant |
| description | Turn ideas into fully formed designs and specs through natural collaborative dialogue. Use when planning new features, designing architecture, or making significant changes to the codebase. |
Feature Design Assistant
Help turn ideas into fully formed designs and specs through structured information gathering and collaborative validation.
Announce at start: "I'm using the feature-design-assistant skill to design this feature."
Phase 1: Context Discovery
First, explore the codebase to understand:
- Project structure and tech stack
- Existing patterns and conventions
- Related features or modules
- Recent changes in relevant areas
Phase 2: Structured Information Gathering
Use AskUserQuestion to batch collect information efficiently. Each call can ask up to 4 questions.
Round 1: Core Requirements (4 questions)
{
"questions": [
{
"question": "What is the primary goal of this feature?",
"header": "Goal",
"multiSelect": false,
"options": [
{ "label": "New Functionality", "description": "Add entirely new capability to the system" },
{ "label": "Enhancement", "description": "Improve or extend existing feature" },
{ "label": "Bug Fix", "description": "Fix incorrect behavior or issue" },
{ "label": "Refactoring", "description": "Improve code quality without changing behavior" }
]
},
{
"question": "Who are the primary users of this feature?",
"header": "Users",
"multiSelect": true,
"options": [
{ "label": "End Users", "description": "External customers using the product" },
{ "label": "Admins", "description": "Internal administrators or operators" },
{ "label": "Developers", "description": "Other developers using APIs or SDKs" },
{ "label": "System", "description": "Automated processes or background jobs" }
]
},
{
"question": "What is the expected scope of this feature?",
"header": "Scope",
"multiSelect": false,
"options": [
{ "label": "Small (1-2 days)", "description": "Single component, limited changes" },
{ "label": "Medium (3-5 days)", "description": "Multiple components, moderate complexity" },
{ "label": "Large (1-2 weeks)", "description": "Cross-cutting concerns, significant changes" },
{ "label": "Unsure", "description": "Need to explore further to estimate" }
]
},
{
"question": "Are there any hard deadlines or constraints?",
"header": "Timeline",
"multiSelect": false,
"options": [
{ "label": "Urgent", "description": "Need this ASAP, within days" },
{ "label": "This Sprint", "description": "Should be done within current sprint" },
{ "label": "Flexible", "description": "No hard deadline, quality over speed" },
{ "label": "Planning Only", "description": "Just designing now, implementing later" }
]
}
]
}
Round 2: Technical Requirements (4 questions)
{
"questions": [
{
"question": "Which layers of the system will this feature touch?",
"header": "Layers",
"multiSelect": true,
"options": [
{ "label": "Data Model", "description": "Database schema, models, migrations" },
{ "label": "Business Logic", "description": "Services, domain logic, rules" },
{ "label": "API", "description": "REST/GraphQL endpoints, contracts" },
{ "label": "UI", "description": "Frontend components, user interface" }
]
},
{
"question": "What are the key quality requirements?",
"header": "Quality",
"multiSelect": true,
"options": [
{ "label": "High Performance", "description": "Must handle high load or be very fast" },
{ "label": "Strong Security", "description": "Sensitive data, auth, access control" },
{ "label": "High Reliability", "description": "Cannot fail, needs redundancy" },
{ "label": "Easy Maintenance", "description": "Needs to be easily understood and modified" }
]
},
{
"question": "How should errors be handled?",
"header": "Errors",
"multiSelect": false,
"options": [
{ "label": "Fail Fast", "description": "Stop immediately on any error" },
{ "label": "Graceful Degrade", "description": "Continue with reduced functionality" },
{ "label": "Retry & Recover", "description": "Automatic retry with recovery logic" },
{ "label": "Context Dependent", "description": "Different strategies for different cases" }
]
},
{
"question": "What testing approach is preferred?",
"header": "Testing",
"multiSelect": false,
"options": [
{ "label": "TDD (Recommended)", "description": "Write tests first, then implementation" },
{ "label": "Test After", "description": "Implement first, add tests after" },
{ "label": "Minimal Tests", "description": "Only critical path testing" },
{ "label": "No Tests", "description": "Skip testing for this feature" }
]
}
]
}
Round 3: Integration & Dependencies (4 questions)
{
"questions": [
{
"question": "Does this feature need external integrations?",
"header": "Integrations",
"multiSelect": true,
"options": [
{ "label": "Database", "description": "New tables, queries, or migrations" },
{ "label": "External APIs", "description": "Third-party service calls" },
{ "label": "Message Queue", "description": "Async processing, events" },
{ "label": "None", "description": "No external integrations needed" }
]
},
{
"question": "Are there dependencies on other features or teams?",
"header": "Dependencies",
"multiSelect": true,
"options": [
{ "label": "Auth System", "description": "User authentication or authorization" },
{ "label": "Other Features", "description": "Depends on features being developed" },
{ "label": "External Team", "description": "Needs input from another team" },
{ "label": "None", "description": "Fully independent feature" }
]
},
{
"question": "How should we handle backwards compatibility?",
"header": "Compat",
"multiSelect": false,
"options": [
{ "label": "Must Maintain", "description": "Cannot break existing clients" },
{ "label": "Version API", "description": "Create new version, deprecate old" },
{ "label": "Breaking OK", "description": "Can make breaking changes" },
{ "label": "Not Applicable", "description": "New feature, no existing users" }
]
},
{
"question": "What documentation is needed?",
"header": "Docs",
"multiSelect": true,
"options": [
{ "label": "API Docs", "description": "Endpoint documentation" },
{ "label": "User Guide", "description": "How-to for end users" },
{ "label": "Dev Guide", "description": "Technical implementation details" },
{ "label": "None", "description": "No documentation needed" }
]
}
]
}
Round 4: Clarifying Questions (Context-Dependent)
Based on previous answers, ask follow-up questions. Examples:
If UI layer selected:
{
"questions": [
{
"question": "What UI framework/approach should we use?",
"header": "UI Tech",
"multiSelect": false,
"options": [
{ "label": "React", "description": "React components with hooks" },
{ "label": "Vue", "description": "Vue.js components" },
{ "label": "Server-Side", "description": "Server-rendered HTML templates" },
{ "label": "Existing Pattern", "description": "Follow current project conventions" }
]
}
]
}
If High Security selected:
{
"questions": [
{
"question": "What security measures are required?",
"header": "Security",
"multiSelect": true,
"options": [
{ "label": "Input Validation", "description": "Strict input sanitization" },
{ "label": "Rate Limiting", "description": "Prevent abuse and DoS" },
{ "label": "Audit Logging", "description": "Track all sensitive actions" },
{ "label": "Encryption", "description": "Encrypt data at rest/transit" }
]
}
]
}
Phase 3: Approach Exploration
After gathering requirements, propose 2-3 approaches:
## Approach Options
### Option A: [Name] (Recommended)
**Pros:** ...
**Cons:** ...
**Best for:** ...
### Option B: [Name]
**Pros:** ...
**Cons:** ...
**Best for:** ...
### Option C: [Name]
**Pros:** ...
**Cons:** ...
**Best for:** ...
Use AskUserQuestion to confirm approach:
{
"questions": [
{
"question": "Which approach would you like to proceed with?",
"header": "Approach",
"multiSelect": false,
"options": [
{ "label": "Option A (Recommended)", "description": "Brief summary of approach A" },
{ "label": "Option B", "description": "Brief summary of approach B" },
{ "label": "Option C", "description": "Brief summary of approach C" }
]
}
]
}
Phase 4: Design Presentation
Present design in sections (300-500 words each), validate after each:
- Architecture Overview - High-level structure
- Data Model - Entities, relationships, schema
- API Design - Endpoints, request/response
- Component Design - Internal modules, interfaces
- Error Handling - Error cases, recovery strategies
- Testing Strategy - What and how to test
After each section, use AskUserQuestion:
{
"questions": [
{
"question": "Does this section look correct?",
"header": "Review",
"multiSelect": false,
"options": [
{ "label": "Looks Good", "description": "Continue to next section" },
{ "label": "Minor Changes", "description": "Small adjustments needed" },
{ "label": "Major Revision", "description": "Significant changes required" },
{ "label": "Questions", "description": "Need clarification before proceeding" }
]
}
]
}
Phase 5: Documentation & Tasks
Save Design Document
Write to docs/designs/YYYY-MM-DD-<topic>-design.md:
# Feature: [Name]
## Summary
[Brief description]
## Requirements
[From Phase 2 answers]
## Architecture
[From Phase 4]
## Implementation Tasks
[Task checklist]
Generate Implementation Tasks
## Implementation Tasks
- [ ] **Task Title** `priority:1` `phase:model` `time:15min`
- files: src/file1.py, tests/test_file1.py
- [ ] Write failing test for X
- [ ] Run test, verify it fails
- [ ] Implement minimal code
- [ ] Run test, verify it passes
- [ ] Commit
- [ ] **Another Task** `priority:2` `phase:api` `deps:Task Title` `time:10min`
- files: src/api.py
- [ ] Write failing test
- [ ] Implement and verify
- [ ] Commit
Phase 6: Execution Handoff
{
"questions": [
{
"question": "How would you like to proceed with implementation?",
"header": "Next Step",
"multiSelect": false,
"options": [
{ "label": "Execute Now", "description": "Run /feature-pipeline in this session" },
{ "label": "New Session", "description": "Start fresh session for implementation" },
{ "label": "Later", "description": "Save design, implement manually later" },
{ "label": "Revise Design", "description": "Go back and modify the design" }
]
}
]
}
Key Principles
- Batch questions efficiently - Use all 4 question slots when appropriate
- Use multiSelect for non-exclusive options - Layers, features, requirements
- Use single-select for decisions - Approach, timeline, strategy
- Mark recommendations - Add "(Recommended)" to preferred options
- Progressive refinement - General → Specific questions
- Validate incrementally - Check understanding at each phase
- YAGNI ruthlessly - Remove unnecessary features from designs