🗓️ 21032026 2100
Primary agent maintains full control and state, using sub-agents as stateless tools
How It Works
- Primary agent owns the entire workflow and state
- Calls sub-agents as stateless functions — they execute a specific task and return results
- Primary agent decides what to do next based on returned results
- Sub-agents have no awareness of the broader workflow
Key Difference from Coordinator
| Coordinator | Agent as Tool | |
|---|---|---|
| Control | Delegates control to sub-agents | Primary agent retains full control |
| State | Sub-agents may manage their own state | Sub-agents are stateless |
| Analogy | Manager delegating work | Craftsman using tools |
When to Use
- Primary agent needs fine-grained control over workflow progression
- Workflow state must be centrally managed (e.g., decisions depend on accumulated context)
- Sub-agent outputs need to be composed or transformed before the next step
Trade-offs
| Pros | Cons |
|---|---|
| Centralized state = easier to reason about overall flow | Primary agent becomes a bottleneck |
| Fine-grained decision making between steps | Similar structural complexity to coordinator |
| Sub-agents are simpler (stateless, single-purpose) | Primary agent prompt can grow complex |
References
- agentic_design_patterns
- coordinator_router_pattern — alternative approach where coordinator delegates control
- Advanced Agentic Patterns for Multi-Agent Systems