Gene: Agent Composition

Description

Compose and combine multiple agents/skills effectively. Enhanced with patterns from agent-skills composition for building complex AI systems.

Trigger Conditions

• User wants to combine multiple capabilities
• Building complex agents from simpler components
• Orchestrating multiple skills
• Workflow automation

Composition Patterns

Sequential Composition

• Chain skills in order
• Output of one becomes input to next
• Maintain context through pipeline

Parallel Composition

• Execute independent skills simultaneously
• Combine results
• Handle partial failures

Conditional Composition

• Route based on conditions
• Branch execution paths
• Select appropriate skills dynamically

Feedback Loops

• Iterative refinement
• Pass output back as input
• Until condition met or max iterations

Design Principles

Modular Design

• Each skill has clear responsibility
• Minimize dependencies between skills
• Make skills independently testable

Abstraction

• Hide implementation details
• Expose clear interfaces
• Compose at appropriate level

Error Handling

• Define error boundaries
• Graceful degradation
• Recovery strategies

Execution Protocol

Step 1: Requirement Analysis

• Identify needed capabilities
• Determine composition pattern
• Map data flow

Step 2: Component Selection

• Select appropriate skills
• Define interfaces
• Plan data transformations

Step 3: Composition

• Wire components together
• Implement data flow
• Handle edge cases

Step 4: Testing & Refinement

• Test each component
• Test composition
• Optimize data flow

Best Practices

• Start simple, add complexity incrementally
• Document composition logic
• Make compositions explicit
• Test failure modes

Integration

• Works with: all other genes
• Foundation for complex agents