Active Agents Programming in HARNESS