Adaptive Agent Glossary

AI is developing so fast, and we hear new vocabulary every day.

Learning the vocabulary of a domain is one of the most efficient ways to learn new things.

Here are some you may be hearing at the moment in adaptive agents, or agents that adapt, or make agents adapt.

Harness: The software scaffolding that surrounds an LLM to turn it into a working agent. It assembles the context the model sees, gives the model the ability to call tools, records results to memory, and re-invokes the model within the agent loop. The LLM supplies the reasoning; the harness supplies the structure around it.

Agent Loop: The repeating cycle by which an agent completes a task: a user request triggers the model to decide on an action, the harness executes a tool and returns the result, and that result is added to the context for the next iteration. The loop continues until the task succeeds or the agent stops.

Trace: A complete record of everything that happened during an agent's run on a task — its conversations, tool calls, decisions, errors, and fixes. Traces are the raw material used to derive skills and other forms of long-term memory.

Working Memory: The messages and tool results currently in use within a single agent run. It exists only for the duration of that run and does not persist afterwards.

Episodic Memory: The record of what happened in previous agent runs. Unlike working memory, episodic memory persists across sessions, allowing an agent to recall specific past interactions rather than just the current one.

Semantic Memory: Project-specific facts and knowledge an agent holds about its environment (for example, the structure or conventions of a codebase). As distinct from episodic memory's record of events, semantic memory stores facts.

Procedural Memory: The workflow steps or reusable methods an agent has learned for accomplishing a task. Where semantic memory captures what is true, procedural memory captures how to do something. This is the form of memory that skills (see Skill Induction) are built from.

Token Space Adaptation: Improving an agent's performance by changing the context and instructions it receives, without altering the underlying model's weights. It covers two methods: behaviour adaptation (skill induction) and knowledge adaptation (code knowledge graphs). Contrast this with Weight Space Adaptation, which modifies the model itself and is generally used only once token space methods are exhausted.

Weight Space Adaptation: Improving an agent's behaviour by directly modifying the weights of the underlying model, typically through fine-tuning. It is more costly than token space adaptation and is used to patch specific behaviours (such as tone or refusals) that context alone cannot reliably produce.