SuperClaude/.claude/commands/spawn.md

Purpose: Parallel specialized agents

---

@include shared/universal-constants.yml#Universal_Legend

Command Execution

Execute: immediate. --plan→show plan first Legend: Generated based on symbols used in command Purpose: "[Action][Subject] in $ARGUMENTS"

Spawn specialized sub-agents for focused tasks with parallel execution capabilities.

@include shared/flag-inheritance.yml#Universal_Always

Examples:

• /spawn --agent researcher "OAuth 2.0 best practices" - Research then implement
• /spawn --mode parallel --agent builder "User auth, Profile API" - Parallel development
• /spawn --mode sequential "Research → Build → Review payment" - Full cycle workflow
• /spawn --mode collaborative --ultrathink "Design microservices" - Collaborative design

Agent Types

Researcher Agent:

• Deep dive into topics
• Compare solutions
• Analyze trade-offs
• Find best practices
• Document findings

Builder Agent:

• Generate code
• Implement features
• Create tests
• Build prototypes
• Integrate systems

Reviewer Agent:

• Code quality checks
• Security analysis
• Performance review
• Best practice validation
• Suggest improvements

Optimizer Agent:

• Performance profiling
• Resource optimization
• Algorithm improvements
• Database tuning
• Cache strategies

Documenter Agent:

• API documentation
• User guides
• Code comments
• Architecture docs
• README files

Execution Modes

Sequential Mode:

Flow: Agent1 → Agent2 → Agent3
Use: When tasks depend on each other
Example: Research → Build → Review

Parallel Mode:

Flow: Agent1 | Agent2 | Agent3
Use: For independent tasks
Example: Multiple feature builds

Collaborative Mode:

Flow: Agents work together
Use: Complex problems
Example: System design session

Best Practices

Task Definition:

• Clear objectives
• Specific deliverables
• Success criteria
• Resource limits
• Time constraints

Agent Selection:

• Match expertise to task
• Consider dependencies
• Plan coordination
• Set boundaries
• Define handoffs

Coordination:

• Clear communication
• Shared context
• Progress tracking
• Result integration
• Quality control

Examples

# Research then implement
/spawn --agent researcher "OAuth 2.0 best practices"
/spawn --agent builder "Implement OAuth based on research"

# Parallel feature development
/spawn --mode parallel --agent builder "User auth, Profile API, Settings UI"

# Full cycle with review
/spawn --mode sequential "Research → Build → Review payment integration"

# Collaborative system design
/spawn --mode collaborative --ultrathink "Design microservices architecture"

Integration

Works with:

• All command flags pass through
• Inherits persona preferences
• Shares project context
• Accumulates findings
• Coordinates outputs

Deliverables

• Agent execution logs
• Task completion reports
• Integrated results
• Performance metrics
• Lessons learned
• Handoff documentation

@include shared/universal-constants.yml#Standard_Messages_Templates