externalization
In plain wordsMoving an AI agent's know-how and memory out of the model itself and into files, tools and rules around it, so the whole system can do more than the model alone.
Agents
The practice and design principle of moving an AI agent's capabilities out of the model's weights and context windowcontext window, n.: How much an AI model can read and keep in view at once: the instructions, the conversation so far and any attached documents. Text that does not fit has to be cut, summarized or left out. into external infrastructure that the model consults and operates: memoryagent memory, n.: The ways an AI assistant or agent keeps information — your preferences, past conversations, lessons from earlier tasks — and brings it back later, beyond what fits in a single conversation. stores that carry state across time, libraries of skillsagent skill, n.: A packaged how-to guide for an AI agent — written steps, sometimes with small programs and examples attached — that sits unread until a job of that kind comes up, when the agent opens it. that hold procedural know-how, protocols such as the Model Context ProtocolModel Context Protocol, prop. n.: An open standard that lets AI assistants plug into outside apps and data — files, calendars, databases, business software — through one common connector instead of a custom link for every pairing; often likened to a USB-C port for AI. that fix how the agent interacts with tools and other agents, and a harnessharness, n.: The software around an AI model that gives it instructions and tools, carries out its actions and decides when it stops. The model is the engine; the harness is the rest of the car. that coordinates these into governed execution. Drawing on the idea of cognitive artifacts — tools such as notebooks and checklists that turn hard mental work into easier tasks — proponents argue that much recent agent progress has come from reorganizing the runtime around a model rather than changing its weights, and frame a trade-off between parametric capability, held in the model, and externalized capability, held in its surroundings.
Instead of fine-tuning, we went for externalization: a skills folder, a memory file and stricter tool schemas.
Origin
Used as the organizing concept of the survey 'Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering' by Chenyu Zhou, Weinan Zhang and 19 co-authors, posted to arXiv on 9 April 2026. The notion of externalized cognition has older roots in cognitive science and the extended-mind thesis.
Related
- harnessharness, n.: The software around an AI model that gives it instructions and tools, carries out its actions and decides when it stops. The model is the engine; the harness is the rest of the car.
- agent memoryagent memory, n.: The ways an AI assistant or agent keeps information — your preferences, past conversations, lessons from earlier tasks — and brings it back later, beyond what fits in a single conversation.
- agent skillagent skill, n.: A packaged how-to guide for an AI agent — written steps, sometimes with small programs and examples attached — that sits unread until a job of that kind comes up, when the agent opens it.
- Model Context ProtocolModel Context Protocol, prop. n.: An open standard that lets AI assistants plug into outside apps and data — files, calendars, databases, business software — through one common connector instead of a custom link for every pairing; often likened to a USB-C port for AI.
- context engineeringcontext engineering, n.: Managing an AI model's reading material: choosing, trimming and arranging the instructions, documents, tool results and memories it is given on every turn, so that it has what the task needs and little else.
- extended mindextended mind, n.: The idea that the mind can reach beyond the head into tools we rely on — a notebook, a phone, perhaps an AI assistant — when they play the same role that memory or reasoning would.
- parametric knowledgeparametric knowledge, n.: What an AI model 'knows' from training, stored in the numbers of its network rather than looked up in documents — like facts you have memorized rather than ones you check in a book.
Sources
- Zhou, C., … Zhang, W. (2026). Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering. arXiv:2604.08224. arxiv.org/abs/2604.08224