Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering
- source_url: https://arxiv.org/abs/2604.08224
- source_type: paper
- platform: arxiv
- author: Chenyu Zhou, Huacan Chai, Wenteng Chen, Zihan Guo, Rong Shan, Yuanyi Song, Tianyi Xu, Yingxuan Yang, Aofan Yu, Weiming Zhang, Congming Zheng, Jiachen Zhu, Zeyu Zheng, Zhuosheng Zhang, Xingyu Lou, Changwang Zhang, Zhihui Fu, Jun Wang, Weiwen Liu, Jianghao Lin, Weinan Zhang
- original_date: 2026-04-09
- added_date: 2026-07-21
- arxiv_id: 2604.08224
- arxiv_categories: cs.SE, cs.MA
- pdf_url: https://arxiv.org/pdf/2604.08224v1
- category: agents
- tags: llm-agents, memory, skills, protocols, harness-engineering, survey, arxiv
- quality_score: 5
摘要（中文）
LLM Agent 的能力正在从“模型内部学会”迁移到“运行时结构外化”：长期记忆、可复用技能、交互协议与 harness 不只是外挂组件，而是在降低模型需要独自承担的认知负担。论文的贡献是把这些工程实践放进统一的 externalization 框架，适合作为整理 agent 基础设施知识图谱的综述入口。
Summary (English)
Large language model (LLM) agents are increasingly built less by changing model weights than by reorganizing the runtime around them. Capabilities that earlier systems expected the model to recover internally are now externalized into memory stores, reusable skills, interaction protocols, and the surrounding harness that makes these modules reliable in practice. This paper reviews that shift through the lens of externalization. Drawing on the idea of cognitive artifacts, we argue that agent infrastructure matters not merely because it adds auxiliary components, but because it transforms hard cognitive burdens into forms that the model can solve more reliably. Under this view, memory externalizes state across time, skills externalize procedural expertise, protocols externalize interaction structure, and harness engineering serves as the unification layer that coordinates them into governed execution. We trace a historical progression from weights to context to harness, analyze memory, skills, and protocols as three distinct but coupled forms of externalization, and examine how they interact inside a larger agent system. We further discuss the trade-off between parametric and externalized capability, identify emerging directions such as self-evolving harnesses and shared agent infrastructure, and discuss open challenges in evaluation, governance, and the long-term co-evolution of models and external infrastructure. The result is a systems-level framework for explaining why practical agent progress increasingly depends not only on stronger models, but on better external cognitive infrastructure.
One-liner
Externalization 不是简单加工具，而是把难题改写成模型更容易可靠求解的系统形态。
原文 / 元数据抓取
Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering
作者: Chenyu Zhou, Huacan Chai, Wenteng Chen, Zihan Guo, Rong Shan, Yuanyi Song, Tianyi Xu, Yingxuan Yang, Aofan Yu, Weiming Zhang, Congming Zheng, Jiachen Zhu, Zeyu Zheng, Zhuosheng Zhang, Xingyu Lou, Changwang Zhang, Zhihui Fu, Jun Wang, Weiwen Liu, Jianghao Lin, Weinan Zhang
原文链接: https://arxiv.org/abs/2604.08224
PDF: https://arxiv.org/pdf/2604.08224v1
发布时间: 2026-04-09
更新时间: 2026-04-09
分类: cs.SE, cs.MA
Large language model (LLM) agents are increasingly built less by changing model weights than by reorganizing the runtime around them. Capabilities that earlier systems expected the model to recover internally are now externalized into memory stores, reusable skills, interaction protocols, and the surrounding harness that makes these modules reliable in practice. This paper reviews that shift through the lens of externalization. Drawing on the idea of cognitive artifacts, we argue that agent infrastructure matters not merely because it adds auxiliary components, but because it transforms hard cognitive burdens into forms that the model can solve more reliably. Under this view, memory externalizes state across time, skills externalize procedural expertise, protocols externalize interaction structure, and harness engineering serves as the unification layer that coordinates them into governed execution. We trace a historical progression from weights to context to harness, analyze memory, skills, and protocols as three distinct but coupled forms of externalization, and examine how they interact inside a larger agent system. We further discuss the trade-off between parametric and externalized capability, identify emerging directions such as self-evolving harnesses and shared agent infrastructure, and discuss open challenges in evaluation, governance, and the long-term co-evolution of models and external infrastructure. The result is a systems-level framework for explaining why practical agent progress increasingly depends not only on stronger models, but on better external cognitive infrastructure.
Obsidian intake evidence excerpt
该内容文件由 AAIF content-fetcher 根据 active/high-score entry 与 OpenCLI arXiv 元数据补齐。
- entry_id: 23d66e0c
- title: Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering
- source: https://arxiv.org/abs/2604.08224
- existing_summary_zh: 综述论文将 LLM agent 的 externalization 归纳为 memoryskillsprotocols 与 harness engineering 等方向，讨论 agent 如何把能力外化到可持久可组合的系统结构中收录理由：它为智能体系统设计提供了统一分类框架，可帮助整理记忆技能协议和执行环境的工程知识