Skills(SKILL.md)は、AIエージェント(Claude Code、Cursor、Codexなど)に特定の能力を追加するための設定ファイルです。
詳しく見る →LLM-as-a-Judge evaluation system using Langfuse. Score AI outputs on relevance, accuracy, hallucination, and helpfulness. Backfill scoring on historical traces. Uses GPT-5-nano for cost-efficient judging. Use when evaluating AI quality, building evals, or monitoring output accuracy.