design-synthesizer

# Architecture Designer output
/docs/implementation/design/architecture-{issue-number}.md

# Documentation Researcher output
/docs/implementation/design/libraries-{issue-number}.md

# Dependency Manager output
/docs/implementation/design/dependencies-{issue-number}.md

### Architecture Summary
**Components**:
- Component A: [Purpose and responsibilities]
- Component B: [Purpose and responsibilities]

**Data Models**:
- Model X: [Fields, validations, relationships]
- Model Y: [Fields, validations, relationships]

**API Contracts**:
- Endpoint 1: [Method, path, request/response]
- Endpoint 2: [Method, path, request/response]

**Data Flow**:
- Flow 1: [Source → Transform → Destination]

**Error Handling**:
- Strategy: [Approach and patterns]

### Library Documentation Summary
**Library A (v1.2.3)**:
- Purpose: [What it does]
- Key APIs: [Main functions/classes]
- Code Example: [Integration pattern]
- Best Practices: [Recommended usage]

**Library B (v2.0.0)**:
- Purpose: [What it does]
- Key APIs: [Main functions/classes]
- Code Example: [Integration pattern]
- Known Issues: [Caveats and workarounds]

### Dependency Summary
**Dependency Tree**:
├── library-a==1.2.3
│   ├── dep-x>=2.0
│   └── dep-y<3.0
└── library-b==2.0.0
    └── dep-z~=1.5

**Conflicts**: [List of version conflicts]
**Resolution**: [How conflicts were resolved]
**Installation Order**: [Sequence of installation]

### Component A
- **Uses**: Library A (v1.2.3)
- **APIs**: Library A's `ClassX` and `FunctionY`
- **Pattern**: Documented in libraries-{issue-number}.md, section 2.3
- **Code Example**: [Link to example in library docs]

### Data Model X
- **Validation**: Uses Library A's `ValidationMixin`
- **Serialization**: Uses Library B's `Serializer`
- **Storage**: Requires dependency: `database-driver>=3.0`

### Library A (v1.2.3)
- **Required Dependencies**: dep-x>=2.0, dep-y<3.0
- **Optional Dependencies**: dep-optional (for feature Z)
- **Compatibility**: Python 3.9+
- **Conflicts**: None identified

**Conflict**: Library A requires dep-x>=2.0, Library B requires dep-x<2.5
**Analysis**: Both constraints can be satisfied with dep-x==2.4.0
**Resolution**: Pin dep-x to 2.4.0 in requirements
**Trade-off**: Cannot use dep-x 2.5+ features until Library B updates
**Risk**: Low - 2.4.0 is stable and well-tested

**Conflict**: Architecture designs async API, but Library C only supports sync
**Analysis**: Need to wrap Library C with async adapter or redesign
**Options**:
  1. Use asyncio.to_thread() to wrap Library C (simpler, slight overhead)
  2. Replace Library C with async alternative (more work, better performance)
**Resolution**: Use asyncio.to_thread() wrapper (priority: speed of implementation)
**Trade-off**: 5-10% performance overhead vs. 2-3 days rewrite
**Mitigation**: Profile performance, optimize if needed

**Conflict**: Architecture uses Library D feature, but no documentation found
**Analysis**: Feature exists in Library D source code but undocumented
**Resolution**:
  1. Add note in PRP about undocumented feature
  2. Include code example from Library D source
  3. Flag for testing during implementation
**Risk**: Medium - undocumented features may change without notice
**Mitigation**: Pin Library D version, add comprehensive tests

### Missing Documentation
- Library E's `advanced_feature()` used in architecture but not documented
- Library F's error handling patterns unclear from documentation

### Incomplete Dependencies
- Component G requires database driver, but version not specified
- Testing framework not included in dependency analysis

### Architecture Gaps
- Component H has no implementation path (needs design detail)
- Error handling for API endpoint 3 not specified

### Unresolved Ambiguities
- Data Model Z validation rules conflict with Library requirements
- Performance target not specified for async operations

## Synthesized Design

### Overview
[High-level description of integrated design]

### Component Architecture
[Unified component diagram with library annotations]

**Component A** (uses Library A v1.2.3)
- **Responsibilities**: [What it does]
- **Implementation**: Uses `Library A.ClassX` for [purpose]
- **Dependencies**: dep-x>=2.0, dep-y<3.0
- **Integration**: [How it connects to other components]

**Component B** (uses Library B v2.0.0)
- **Responsibilities**: [What it does]
- **Implementation**: Uses `Library B.ServiceY` for [purpose]
- **Dependencies**: dep-z~=1.5
- **Integration**: [How it connects to other components]

### Data Models (Pydantic Schemas)

**UserProfile Model**
```python
from library_a import ValidationMixin
from pydantic import BaseModel

class UserProfile(BaseModel, ValidationMixin):
    id: int
    name: str
    email: EmailStr  # From library_a
    # ... fields with library-specific validations

library-a==1.2.3
├── dep-x==2.4.0  # Pinned to resolve conflict
├── dep-y==2.8.0
library-b==2.0.0
├── dep-z==1.5.3
└── dep-x==2.4.0  # Shared with library-a

pip install library-a==1.2.3
pip install library-b==2.0.0
pip install -r requirements.txt

### Step 8: Validate Synthesis Quality

Quality checklist:

**Completeness**:
- ✅ All architecture components mapped to libraries
- ✅ All libraries mapped to dependencies
- ✅ All data models have implementation details
- ✅ All API contracts have library integration notes
- ⚠️  Component H needs more detail

**Consistency**:
- ✅ Terminology consistent across documents
- ✅ Version numbers match everywhere
- ✅ Component names standardized
- ✅ Dependencies resolved without conflicts

**Actionability**:
- ✅ Implementation path is clear and specific
- ✅ Installation commands provided
- ✅ Code patterns documented
- ⚠️  Some gaps flagged for resolution

**Traceability**:
- ✅ Design elements link to requirements
- ✅ Library choices justified
- ✅ Design decisions documented with rationale

## Best Practices

### 1. Start with Overview
- Read all outputs completely before synthesis
- Create summary of each output
- Identify major themes and patterns

### 2. Map Systematically
- Use structured templates for mapping
- Create traceability matrices
- Document every cross-reference

### 3. Validate Early
- Check consistency as you integrate
- Flag issues immediately
- Don't ignore small discrepancies

### 4. Resolve Conflicts Transparently
- Document all conflicts found
- Explain resolution reasoning
- Note trade-offs and risks

### 5. Identify All Gaps
- Missing documentation
- Incomplete specifications
- Unresolved ambiguities
- Assumptions that need validation

### 6. Create Unified Narrative
- Tell coherent story from architecture → libraries → dependencies → implementation
- Link every design element to requirements
- Provide clear implementation path

## Resources

### synthesis-guide.md
Comprehensive guide for synthesizing sub-agent outputs including:
- Cross-referencing methodology
- Consistency checking procedures
- Conflict resolution strategies
- Gap identification techniques
- Quality validation criteria

### design-patterns.md
Common design patterns and structures including:
- Component integration patterns
- Library usage patterns
- Dependency management patterns
- Error handling patterns
- Testing patterns

## Example Usage

### Input (Sub-Agent Outputs)

**Architecture Designer Output**:
```markdown
## Components
- **UserService**: Handles user CRUD operations
- **TaskQueue**: Manages background tasks

## Data Models
- **User**: id, name, email, created_at
- **Task**: id, task_type, status, payload

## APIs
- POST /api/users: Create user
- POST /api/tasks: Queue background task

## Libraries
**SQLAlchemy (v2.0.0)**: ORM for database
- User model: Use `declarative_base()` pattern
- Example: [Code snippet]

**Celery (v5.3.0)**: Background task queue
- Task definition: Use `@app.task` decorator
- Example: [Code snippet]

## Dependencies
sqlalchemy==2.0.0
├── greenlet>=2.0
celery==5.3.0
├── kombu>=5.3
└── billiard>=4.0

Conflict: greenlet version (resolved to 2.0.2)

## Synthesized Design

### Component Architecture

**UserService** (uses SQLAlchemy v2.0.0)
- **Implementation**: SQLAlchemy `declarative_base()` pattern
- **Dependencies**: sqlalchemy==2.0.0, greenlet==2.0.2
- **Data Model**: User class inherits from SQLAlchemy Base

**TaskQueue** (uses Celery v5.3.0)
- **Implementation**: Celery `@app.task` decorator
- **Dependencies**: celery==5.3.0, kombu>=5.3, billiard>=4.0
- **Integration**: Tasks defined in tasks.py, queued via TaskQueue service

### Data Models (SQLAlchemy + Pydantic)

```python
from sqlalchemy import declarative_base
from pydantic import BaseModel

Base = declarative_base()

class User(Base):
    __tablename__ = 'users'
    id: int
    name: str
    email: str  # With SQLAlchemy validation
    created_at: datetime

sqlalchemy==2.0.0
├── greenlet==2.0.2  # Pinned to resolve conflict
celery==5.3.0
├── kombu>=5.3
└── billiard>=4.0

pip install sqlalchemy==2.0.0 celery==5.3.0

## Integration

This skill is used by:
- **design-orchestrator** agent during Phase 2: Design & Planning
- Activates automatically when orchestrator synthesizes sub-agent outputs
- Provides unified design for PRP generation (prp-generator skill)

---

**Version**: 2.0.0
**Auto-Activation**: Yes (when synthesizing sub-agent outputs)
**Phase**: 2 (Design & Planning)
**Created**: 2025-10-29