A method for constructing an agent in an execution environment that
utilizes a set of supplied canonical rules and a set of environment
resident canonical rules comprising retrieving a first canonical rule set
for the agent, retrieving a second canonical rule set for the agent and
merging the two rule sets.