Outcomes

A self-supervised agent execution mechanism in which the user defines the task, the evaluation criteria and the maximum number of iterations. After the agent executes, a verification sub-agent checks the output against the criteria and retries until the requirements are met.