Back to Home
arXiv AI··Papers & Tech

Agent Harness Awareness: Turning Verified Telemetry into Operational Control for Tool-Using Agents

中文摘要

介绍 Agent Harness Awareness 设计,通过实时遥测信息消除智能体的“运行盲区”,使其能根据工具的实际运行状态做出决策。

English Summary

Agent Harness Awareness bridges the "operational blindness" gap by providing tool-using agents with real-time, verified telemetry to assess a tool's current health and suitability.

Original Excerpt

arXiv:2611.05391v1 Announce Type: new Abstract: Tool descriptions tell an agent what an action can do, but they do not necessarily reveal whether that action is healthy or suitable at the moment of use. Execution harnesses often possess this information through monitors, traces, and recent failures while the model choosing the next action does not. We call this gap operational blindness. We introduce Agent Harness Awareness, a trust-separated design in which the host supplies a small, fresh, verified state snapshot before a consequential action while retaining measurement, enforcement, evaluation, and final dispatch authority. We test its tools-and-network instance with ToolRoute, a controlled isolation benchmark that holds the task and candidate semantics constant and varies the operational information available to the model. In a 216-decision ablation across GPT-5.6 Sol, Claude Sonnet 5, and Gemini 3.7 Flash, verified tool-health state reduced mean observable policy regret by 96.9% against a same-shape neutral control (4,252 ms to 131 ms) and raised single-dispatch completion from 69.4% to 95.8%. In a separate 27-decision end-to-end replication with live local TCP/HTTP execution,…