[Feature Request]: ReasoningAgent Interim Execution of steps

[giorgossideris]

Is your feature request related to a problem? Please describe.

Enhancing the deep thinking and reasoning capabilities of ReasoningAgent would benefit from a more dynamic approach. Instead of merely suggesting the next step at each stage and executing the entire process at once, it would be more effective to allow interim execution of individual steps. This would enable the agent to generate new suggestions based on the outcomes of each executed step, fostering a more adaptive and iterative reasoning process.

Describe the solution you'd like

An option for interim execution of steps

Additional context

No response