agent-evaluation

【开发与工程】ALWAYS use this when the request matches Agent Evaluation: Testing and benchmarking LLM agents including behavioral testing, capability assessment, reliability metrics, and production monitoring—where even top agents achieve less than 50% on real-world benchmarks

独立 GitHub 快照 · 自动分类 · 以原作者文档为准