Currently it is not really possible for agents to request help and then complete a behaviour with another agent. This could be useful for cases where a treatment etc requires multiple staff members to complete (e.g. escorting a disabled patient to a ward, a nurse requiring help administering a test etc)
This could, again, be a case where it is useful to split the modelling at the domain-modelling and agent-modelling levels:
At the domain-modelling level, this may simply be an annotation to an activity (in an action card?) stating that multiple staff are needed for the activity.
At the agent-modelling level, this would then be translated into a specific algorithm describing how agents go about finding the people they need for an activity.
At the domain-modelling level, there may eventually be a separate small modelling language that allows the description of different strategies agents might employ for finding partners to work with. This would allow configuring different strategies for different activities or different models.
Currently it is not really possible for agents to request help and then complete a behaviour with another agent. This could be useful for cases where a treatment etc requires multiple staff members to complete (e.g. escorting a disabled patient to a ward, a nurse requiring help administering a test etc)