← Concept IndexDEFINITION WHY IT MATTERS COMMONLY CONFUSED WITH
Prompting & context engineering
Breaking work into steps
Also called: task decomposition, one step at a time
Splitting a big request into a sequence of smaller, checkable steps — outline, then draft, then critique — instead of asking for the whole thing in one shot.
Smaller steps are easier for the model to do well and easier for you to verify, catching a wrong turn early instead of at the end. It's the everyday form of 'chain-of-thought' you control.
Chain-of-thought inside the model. Here you own the steps and inspect each one, rather than trusting a single generated explanation.