It is commonly believed there is a big gap between agent logics and computational agent frameworks. In this paper, we show that this gap is not as big as believed by showing that GOAL agents instantiate Intention Logic of Cohen and Levesque. That is, we show that GOAL agent programs can be formally related to Intention Logic. We do so by proving that the GOAL Verification Logic can be embedded into Intention Logic. It follows that (a fragment of) Intention Logic can be used to prove properties of GOAL agents. The work reported is an important step towards the application of standard tools from modal logic for e.g. model checking agent programs. Our results also prove useful for extending the expressiveness of the GOAL agent language. This is illustrated by incorporating temporally extended goals into GOAL agents.