Execution modes
NucleusIQ uses the Gearbox Strategy—three execution modes that scale from simple chat to autonomous reasoning.
Comparison
| Capability | Direct | Standard | Autonomous |
|---|---|---|---|
| Memory | Yes | Yes | Yes |
| Plugins | Yes | Yes | Yes |
| Tools | Yes (max 5) | Yes (max 30) | Yes (max 100) |
| Tool loop | Yes | Yes | Yes |
| Task decomposition | No | No | Yes |
| Critic (verification) | No | No | Yes |
| Refiner (correction) | No | No | Yes |
| Validation pipeline | No | No | Yes |
Tool limits are configurable via AgentConfig(max_tool_calls=N).
Direct
Fast, single LLM call. Best for Q&A and simple lookups.
agent = Agent(..., config=AgentConfig(execution_mode=ExecutionMode.DIRECT))
Standard
Tool-enabled, linear execution. Default for most workflows.
agent = Agent(..., config=AgentConfig(execution_mode=ExecutionMode.STANDARD))
# or omit — STANDARD is the default
Autonomous
Planning, multi-step execution with Critic/Refiner verification. For complex, high-stakes tasks.
agent = Agent(..., config=AgentConfig(execution_mode=ExecutionMode.AUTONOMOUS))