Agent harness

Rare card · No. 47490

An agent harness, also known as agent scaffolding, is the software infrastructure surrounding a large language model (LLM) that enables it to operate as an AI agent. It manages tool use, memory, state persistence, execution environments and feedback loops, as opposed to the model's internal reasoning. The UK's AI Security Institute described an AI agent as being formed by the model plus the scaffolding back in 2023. The relationship can be expressed as agent = model + harness. Because a "base model" LLM is a stateless probabilistic function that maps input token sequences to output token sequences, the harness is what makes an agent out of a model, i.e. a thing that interfaces with external tools, inputs and outputs non-token data (such as images), and interacts with a task-environment coherently over many rounds. Rather than repeatedly re-reading an ever-growing transcript inside the context window, a harness can offload record-keeping into a structured software environment that manages...

Read on Wikipedia · Image Wikideas1, CC0