One-shot learning can be achieved by animals and algorithms, but how animals do it is poorly understood as most of the algorithms are not biologically plausible. Experiments studying one-shot learning in rodents have shown that after initial gradual learning of associations between cues and locations, new associations can be learned with just a single exposure to each new cue-location pair. Foster, Morris and Dayan (2000) developed a neuro-symbolic actor-critic and coordinate learning agent that exhibited one-shot learning to displaced single locations in an open field maze using dead reckoning. While the temporal difference rule for learning the agent’s coordinates was biologically plausible, the agent’s memory mechanism for learning target coordinates was not, nor did they address one-shot learning of multiple cue-location pairs that rodents are also capable of (Tse et al., 2007). Here we extend the biological plausibility of that agent by replacing the symbolic memory mechanism with a reservoir of recurrently connected neurons resembling cortical microcircuitry. Biologically plausible learning of goal coordinates was achieved by subjecting the reservoir’s output weights to synaptic plasticity governed by a novel 4-factor variant of the exploratory Hebbian (EH) rule gated by reward. The agent’s current coordinates and goal coordinates were passed to a pretrained neural network that performed vector subtraction and selected the direction of movement towards the target. Our fully neural agent trained by Hebbian plasticity combines functions thought to involve the hippocampus and prefrontal cortex such that the memory system can store in one shot goal coordinates that can be recalled when a relevant cue is presented, while the coordinate system acts as a cognitive map encoding relational information for goal directed dead reckoning. As with rodents, the biologically plausible agent exhibited one-shot learning in the multiple cue-location paired associations task of Tse and colleagues.