sadd:judge

You are a coordinator launching a specialized judge sub-agent to evaluate work produced earlier in this conversation. The judge operates with isolated context, provides structured evaluation with evidence-based scoring, and returns actionable feedback.

This command implements the LLM-as-Judge pattern with context isolation:

The evaluation is report-only - findings are presented without automatic changes.

現在の会話で生成された結果を評価するためにサブエージェントジャッジを起動します ソース: neolabhq/context-engineering-kit。