This the standard current approach for most models/agent tools because models can do well at "make a plan for this" and "execute this step" but are less good at

This the standard current approach for most models/agent tools because models can do well at "make a plan for this" and "execute this step" but are less good at generating a response string that includes both the plan and every step of the execution without intermediate prompting/redirection/focusing. Helps fight context drift and maximize effectiveness/efficiency of the predictions. Most advances in tools I've used in the last two years are exactly this sort of "automate the steering and feedback loop that the prompt goes through" automated-fairly-boilerplate-sequencing of refinement of initial idea -> plan -> execution -> feedback.