Planner-executor architecture

An agent design pattern where one component decides what steps to take (the planner) and a separate component carries them out (the executor), making it easier to debug, adjust, and improve each part independently.

In this architecture, the planner looks at the goal, breaks it into steps, and produces a plan. The executor takes each step and performs the actual action: calling a tool, querying a database, or writing a file. After each action, the result goes back to the planner, which can revise the remaining steps based on what happened. For example, a research agent's planner might outline "search for X, read the top three results, summarize findings," while the executor handles the actual search calls and file reads. The planner and executor can be separate models, separate prompts within one model, or a mix of model and code.

Builder example

This separation gives you a clear diagnostic surface. When something goes wrong, you can check the plan to see if the steps were sensible, then check the execution to see if each step ran correctly. Without this separation, debugging a failed multi-step task means untangling reasoning and action in a single stream of output, which is much harder.

Common confusion: The planner is not inherently smarter or more reliable than the executor. A planner with poor context will produce bad plans, and a good plan can still fail if the executor lacks proper error handling. Both layers need attention.