Repository navigation
02 personas persona selection guide
This guide helps you choose the right Coherence APM Framework v4.2.0 agent for your current task. Each of the 11 specialized agents is designed for specific aspects of software development and project management.
What do you need to do?
│
├─ 💡 "I have an idea to explore"
│ └─ Use: /analyst
│
├─ 📄 "I need to document requirements"
│ └─ Use: /pm (Product Manager)
│
├─ 🏗️ "I need to design the system"
│ └─ Use: /architect
│
├─ 🎨 "I need to design the UI/UX"
│ └─ Use: /design-architect
│
├─ 📋 "I need to manage the backlog"
│ └─ Use: /po (Product Owner)
│
├─ 📝 "I need to create user stories"
│ └─ Use: /sm (Scrum Master)
│
├─ 💻 "I need to write code"
│ └─ Use: /dev (Developer) or /developer (Senior Developer)
│
├─ 🔍 "I need to test/validate"
│ └─ Use: /qa (Quality Assurance)
│
└─ 🎯 "I need coordination/guidance"
└─ Use: /coherence (Coherence Orchestrator)
Command: /coherence
When to Use:
- Starting a new session
- Need guidance on which agent to use
- Coordinating multiple agents
- Managing complex workflows
- Getting project status updates
- Unified context engineering
Specialties:
- Unified context engineering
- Native sub-agent coordination
- Central workflow management
- Agent delegation with seamless handoffs
- Context preservation and session management
- Project oversight with orchestrated intelligence
Voice Integration: Uses speakOrchestrator.sh with configurable TTS providers
Example Scenarios:
"Help me plan my project workflow"
"Which agent should I use for this task?"
"Show me the project status"
"Guide me through building an API"
"Initialize unified context engineering"
Command: /coherence
When to Use:
- Legacy system compatibility
- Backward compatibility requirements
Note: Consider migrating to /coherence for enhanced capabilities
Command: /analyst
When to Use:
- Starting a new project
- Exploring ideas
- Gathering requirements
- Researching solutions
- Validating concepts
Specialties:
- Brainstorming facilitation
- Requirements gathering
- Market research
- Feasibility analysis
- Problem definition
Voice Integration: Uses speakAnalyst.sh with TTS support
Example Scenarios:
"Help me brainstorm a fitness tracking app"
"Research authentication best practices"
"Analyze requirements for a payment system"
"Validate this business idea"
Parallel Alternative: /planning-brainstorming
Command: /pm
When to Use:
- Creating Product Requirements Documents (PRD)
- Defining product vision
- Planning features
- Setting success metrics
- Stakeholder communication prep
Specialties:
- PRD creation
- Feature prioritization
- User story mapping
- Success metrics definition
- Roadmap planning
Voice Integration: Uses speakPm.sh with TTS support
Example Scenarios:
"Create a PRD for our todo application"
"Define success metrics for this feature"
"Prioritize these feature requests"
"Plan the product roadmap"
Parallel Alternative: /planning-prd (70% faster)
Command: /architect
When to Use:
- Designing system architecture
- Choosing technology stack
- Planning integrations
- Defining APIs
- Solving technical challenges
Specialties:
- System design
- Technology selection
- Architecture patterns
- Scalability planning
- Security architecture
Voice Integration: Uses speakArchitect.sh with TTS support
Example Scenarios:
"Design architecture for a chat application"
"Choose the tech stack for our project"
"Plan microservices architecture"
"Design the Coherence systemI structure"
Parallel Alternative: /planning-architecture (75% faster)
Command: /design-architect
When to Use:
- Designing user interfaces
- Creating UX specifications
- Planning frontend architecture
- Defining design systems
- Creating UI component libraries
Specialties:
- UI/UX design
- Frontend architecture
- Component design
- Design systems
- Responsive layouts
Voice Integration: Uses speakDesignArchitect.sh with TTS support
Example Scenarios:
"Design the UI for our dashboard"
"Create a design system"
"Plan the frontend architecture"
"Design mobile-responsive layouts"
Parallel Alternative: /planning-frontend-architecture
Command: /po
When to Use:
- Managing product backlog
- Grooming user stories
- Planning sprints
- Prioritizing work
- Defining acceptance criteria
Specialties:
- Backlog management
- Story grooming
- Sprint planning
- Prioritization
- Stakeholder management
Voice Integration: Uses speakPo.sh with TTS support
Example Scenarios:
"Groom the backlog for next sprint"
"Create epics from the PRD"
"Prioritize these user stories"
"Define acceptance criteria"
Parallel Alternative: /planning-epic, /planning-stories
Command: /sm
When to Use:
- Creating user stories
- Planning sprints
- Facilitating ceremonies
- Removing blockers
- Process improvement
Specialties:
- Story creation
- Sprint facilitation
- Process optimization
- Team coordination
- Agile coaching
Voice Integration: Uses speakSm.sh with TTS support
Example Scenarios:
"Create user stories for authentication"
"Plan the next sprint"
"Help resolve this blocker"
"Optimize our development process"
Parallel Alternative: /planning-next-story
Command: /dev
When to Use:
- Writing code
- Implementing features
- Debugging issues
- Basic code refactoring
- Standard technical implementation
Specialties:
- Code implementation
- Basic algorithm design
- Debugging
- Unit testing
- Code review
Voice Integration: Uses speakDeveloper.sh with TTS support
Example Scenarios:
"Implement user authentication"
"Debug this error"
"Write unit tests"
"Implement this feature"
Command: /developer
When to Use:
- Complex code architecture
- Advanced debugging
- Performance optimization
- Code refactoring for scalability
- Technical leadership tasks
Specialties:
- Advanced architecture patterns
- Performance optimization
- Complex algorithm design
- Code refactoring
- Technical mentoring
Voice Integration: Uses speakDeveloper.sh with TTS support
Example Scenarios:
"Optimize this algorithm for performance"
"Refactor this code for better architecture"
"Design a complex data structure"
"Lead technical implementation"
Note: Both developer personas can spawn multiple instances for parallel development
Command: /qa
When to Use:
- Creating test plans
- Writing test cases
- Performing testing
- Quality validation
- Bug tracking
Specialties:
- Test strategy
- Test automation
- Quality metrics
- Bug detection (92-
- Performance testing
Voice Integration: Uses speakQa.sh with TTS support
Example Scenarios:
"Create test plan for this feature"
"Write test cases for authentication"
"Perform security testing"
"Validate the implementation"
Parallel Alternative: /qa-framework (comprehensive testing)
1. /coherence → Initialize Coherence Orchestrator
2. /analyst → Explore and research
3. /pm → Create PRD
4. /architect → Design system
5. /po → Create backlog
6. /sm → Generate stories
7. /dev or /developer → Implement
8. /qa → Test and validate
1. /analyst → Research requirements
2. /architect → Technical design
3. /sm → Create stories
4. /dev → Implementation
5. /qa → Testing
1. /analyst → Quick ideation
2. /design-architect → UI mockup
3. /dev → Rapid prototype
1. /qa → Identify issue
2. /dev → Fix implementation
3. /qa → Validate fix
Use parallel commands when you need:
- Multiple analyses simultaneously
- Faster document generation
- Batch processing of similar tasks
- Comprehensive coverage quickly
| Sequential | Parallel | Speed Gain |
|---|---|---|
/analyst |
/planning-brainstorming |
4x |
/pm |
/planning-prd |
70% faster |
/architect |
/planning-architecture |
75% faster |
/po + /sm
|
/planning-stories |
4x |
/qa |
/qa-framework |
4x |
Always begin with /coherence to initialize the unified context engineering system and get guidance.
Idea → Research → Requirements → Design → Implementation → Testing
Each agent is an expert - use them for their specialty.
When time matters, use parallel commands.
Agents share session notes - no need to repeat information.
| I Need To... | Best Agent | Alternative | Parallel Option |
|---|---|---|---|
| Explore ideas | Analyst | PM | /planning-brainstorming |
| Write requirements | PM | Analyst | /planning-prd |
| Design system | Architect | Design Arch | /planning-architecture |
| Design UI | Design Arch | Architect | /planning-frontend-architecture |
| Manage backlog | PO | SM | /planning-stories |
| Create stories | SM | PO | /planning-next-story |
| Write code | Developer/Senior Dev | - | Multiple developers |
| Test code | QA | Developer | /qa-framework |
| Coordinate | Coherence Orchestrator | Coherence Orchestrator (legacy) | - |
Analyst → PM → Architect → Developer → QA
Each agent completes their work before handing off.
┌→ Developer 1
Architect → Developer 2
└→ Developer 3
Multiple agents work simultaneously.
PM ↔ Architect ↔ Developer
Agents collaborate back and forth.
Analyst
↓
Orchestrator → PM
↓ ↘ Architect
Developer ↘ QA
Orchestrator coordinates multiple specialists.
- Wrong: Using Developer for requirements gathering
- Right: Use Analyst or PM for requirements
- Wrong: Jumping straight to development
- Right: Research → Plan → Design → Develop
- Wrong: Sequential execution for everything
- Right: Use parallel for appropriate tasks
- Wrong: Repeating information to each agent
- Right: Agents share context automatically
/planning-brainstorming # Quick ideation
/design-architect # Fast UI design
/dev # Immediate implementation/planning-prd # Complete PRD
/planning-architecture # Full system design
/planning-stories # All user stories/qa-framework # Set quality standards first
/dev # Develop with quality in mind
/parallel-test # Comprehensive testing-
Agent Details: See individual guides in
../02-personas/ -
Workflows: Check
../03-workflows/ -
Commands: Reference
../04-commands/ -
Examples: Browse
../examples/
- Can't decide? Use
/coherenceand ask for guidance - Want details? Read the specific agent guide
- Need examples? Check the examples directory
- Still stuck? See troubleshooting guide
Remember: Each agent is a specialist. Use the right expert for the right job!