BMAD-METHOD/.bmad-core/tasks/create-team.md

Create Team Task

This task guides you through creating a new BMAD agent team that conforms to the agent-team schema and effectively combines agents for specific project types.

Note for User-Created Teams: If creating a custom team for your own use (not part of the core BMAD system), prefix the team name with a period (e.g., .team-frontend) to ensure it's gitignored and won't conflict with repository updates.

Prerequisites

1. Load and understand the team schema: /bmad-core/schemas/agent-team-schema.yml
2. Review existing teams in /bmad-core/agent-teams/ for patterns and naming conventions
3. List available agents from /agents/ to understand team composition options
4. Review workflows in /bmad-core/workflows/ to align team capabilities

Process

1. Define Team Purpose and Scope

Before selecting agents, clarify the team's mission:

• Team Purpose: What specific problems will this team solve?
• Project Types: Greenfield, brownfield, or both?
• Technical Scope: UI-focused, backend-only, or full-stack?
• Team Size Consideration: Smaller teams (3-5 agents) for focused work, larger teams (6-8) for comprehensive coverage

2. Create Team Metadata

Based on the schema requirements:

• Team Name: Must follow pattern ^Team .+$ (e.g., "Team Frontend", "Team Analytics")
  • For user teams: prefix with period (e.g., "Team .MyCustom")
• Description: 20-500 characters explaining team's purpose, capabilities, and use cases
• File Name: /bmad-core/agent-teams/team-{identifier}.yml
  • For user teams: /bmad-core/agent-teams/.team-{identifier}.yml

3. Select Agents Based on Purpose

Discover Available Agents

1. List all agents from /agents/ directory
2. Review each agent's role and capabilities
3. Consider agent synergies and coverage

Agent Selection Guidelines

Based on team purpose, recommend agents:

For Planning & Strategy Teams:

• bmad (required orchestrator)
• analyst - Requirements gathering and research
• pm - Product strategy and documentation
• po - Validation and approval
• architect - Technical planning (if technical planning needed)

For Design & UX Teams:

• bmad (required orchestrator)
• ux-expert - User experience design
• architect - Frontend architecture
• pm - Product requirements alignment
• po - Design validation

For Development Teams:

• bmad-orchestrator (required)
• sm - Sprint coordination
• dev - Implementation
• qa - Quality assurance
• architect - Technical guidance

For Full-Stack Teams:

• bmad-orchestrator (required)
• analyst - Initial planning
• pm - Product management
• ux-expert - UI/UX design (if UI work included)
• architect - System architecture
• po - Validation
• Additional agents as needed

Special Cases

• Using Wildcard: If team needs all agents, use ["bmad", "*"]
• Validation: Schema requires bmad in all teams

4. Select Workflows

Based on the schema's workflow enum values and team composition:

1. Analyze team capabilities against available workflows:

   • brownfield-fullstack - Requires full team with UX
   • brownfield-service - Backend-focused team
   • brownfield-ui - UI/UX-focused team
   • greenfield-fullstack - Full team for new projects
   • greenfield-service - Backend team for new services
   • greenfield-ui - Frontend team for new UIs

2. Match workflows to agents:

   • UI workflows require ux-expert
   • Service workflows benefit from architect and dev
   • All workflows benefit from planning agents (analyst, pm)

3. Apply schema validation rules:

   • Teams without ux-expert shouldn't have UI workflows
   • Teams named "Team No UI" can't have UI workflows

5. Create Team Configuration

Generate the configuration following the schema:

bundle:
  name: "{Team Name}" # Must match pattern "^Team .+$"
  description: >-
    {20-500 character description explaining purpose,
    capabilities, and ideal use cases}

agents:
  - bmad # Required orchestrator
  - { agent-id-1 }
  - { agent-id-2 }
  # ... additional agents

workflows:
  - { workflow-1 } # From enum list
  - { workflow-2 }
  # ... additional workflows

6. Validate Team Composition

Before finalizing, verify:

1. Role Coverage: Does the team have all necessary skills for its workflows?
2. Size Optimization:
   • Minimum: 2 agents (bmad + 1)
   • Recommended: 3-7 agents
   • Maximum with wildcard: bmad + "*"
3. Workflow Alignment: Can the selected agents execute all workflows?
4. Schema Compliance: Configuration matches all schema requirements

7. Integration Recommendations

Document how this team integrates with existing system:

1. Complementary Teams: Which existing teams complement this one?
2. Handoff Points: Where does this team hand off to others?
3. Use Case Scenarios: Specific project types ideal for this team

8. Validation and Testing

1. Schema Validation: Ensure configuration matches agent-team-schema.yml
2. Build Validation: Run npm run validate
3. Build Team: Run npm run build:team -t {team-name}
4. Size Check: Verify output is appropriate for target platform
5. Test Scenarios: Run sample workflows with the team

Example Team Creation

Example 1: API Development Team

bundle:
  name: "Team API"
  description: >-
    Specialized team for API and backend service development. Focuses on
    robust service architecture, implementation, and testing without UI
    components. Ideal for microservices, REST APIs, and backend systems.

agents:
  - bmad
  - analyst
  - architect
  - dev
  - qa
  - po

workflows:
  - greenfield-service
  - brownfield-service

Example 2: Rapid Prototyping Team

bundle:
  name: "Team Prototype"
  description: >-
    Agile team for rapid prototyping and proof of concept development.
    Combines planning, design, and implementation for quick iterations
    on new ideas and experimental features.

agents:
  - bmad
  - pm
  - ux-expert
  - architect
  - dev

workflows:
  - greenfield-ui
  - greenfield-fullstack

Team Creation Checklist

• Team purpose clearly defined
• Name follows schema pattern "Team {Name}"
• Description is 20-500 characters
• Includes bmad orchestrator
• Agents align with team purpose
• Workflows match team capabilities
• No conflicting validations (e.g., no-UI team with UI workflows)
• Configuration validates against schema
• Build completes successfully
• Output size appropriate for platform

Best Practices

1. Start Focused: Create teams with specific purposes rather than general-purpose teams
2. Consider Workflow: Order agents by typical workflow sequence
3. Avoid Redundancy: Don't duplicate roles unless needed
4. Document Rationale: Explain why each agent is included
5. Test Integration: Verify team works well with selected workflows
6. Iterate: Refine team composition based on usage

This schema-driven approach ensures teams are well-structured, purposeful, and integrate seamlessly with the BMAD ecosystem.