用 NVIDIA NeMo Relay 追踪 Agent Harness 行为
原标题：Tracing Agent Harness Behavior with NVIDIA NeMo Relay
AI 摘要
NVIDIA 发布技术博客教程，介绍如何用 NVIDIA NeMo Relay 追踪 Hermes Agent 的行为。教程通过两个示例任务，展示如何利用 ATOF、ATIF 和 OpenTelemetry 三种格式的轨迹数据，检查模型与工具调用、错误、重试、耗时和 token 使用，并结合任务验证结果评估 agent harness 的改动。NeMo Relay 为 agent 可观测性和安全治理提供结构化证据层。
正文节选
An agent can finish a task and still take an inefficient path. A failed search can trigger another search. A truncated file read can lead to a command fetching the same content again. A correct final answer hides those extra steps, even though they increase latency and consume tokens. Inefficiencies create more chances for failure. To improve an agent’s behavior, developers must understand whether a task succeeded and how the agent completed it. A success check by itself cannot explain why an ag