← Concept Index

Agents — core

Memory & state

Also called: what the agent remembers, scratchpad, carrying context

DEFINITION

How an agent keeps track of what it has done and learned across steps — from a running scratchpad within one task to stored memory that persists across sessions.

WHY IT MATTERS

Memory is what makes an agent coherent over a long task, but it's also a risk surface: stored context can hold sensitive data, go stale, or carry a mistake (or an injected instruction) forward into later actions.

COMMONLY CONFUSED WITH

The model being retrained. Agent memory lives outside the weights — it's context the system stores and re-feeds — not permanent learning by the model.