Multi-agent methodologies’ incoherencies