用 NVIDIA NeMo 将智能体失败转化为测试并修复
原标题：Use NVIDIA NeMo to Turn Agent Failures into Tests and Harness Fixes
AI 摘要
NVIDIA 发布技术博客，介绍如何用 NVIDIA NeMo 工具链把智能体在生产环境中的失败案例转化为测试用例并修复。流程包括用 NeMo Compass 分析 42 条 Hermes 智能体轨迹，发现其在未获用户确认时调用 send_message 工具的问题；再用 NeMo Eval Author 生成 Harbor 评估与验证器复现该失败；最后修改智能体指令（SOUL.md）并在开发集与留出任务上重新评估。文章称该流程可帮助改进提示、工具、模型、记忆与 harness 代码。
正文节选
Turn production evidence into prioritized issues and custom evaluations that help you improve prompts, tools, models, memory, skills, and harness code. When you build an agent for your domain, generic benchmark scores may not explain its failures. An agent might choose the wrong tool, follow the wrong workflow branch, miss relevant context, use an unnecessarily expensive model, or complete a task without meeting a domain-specific requirement. Improving an agent requires more than model tuning. P