do-and-judge

Execute a task with sub-agent implementation and LLM-as-a-judge verification with automatic retry loop

Independent GitHub snapshot · Automated classification · Original documentation governs