# Introduction

NEW IN 2026

Harness engineering emerged through 2026 as a distinct discipline, formalized in surveys covering 100+ papers and 23 production systems. The core claim: at fixed model capability, agent-computer interface design materially changes benchmark outcomes.

## The Equation:

##  Agent = Model + Harness

The harness treats the LLM as a frozen reasoning utility and moves the responsibility for safety, execution accuracy, multi-step orchestration, and adaptive memory into the surrounding infrastructure. A production harness is generally described as a three-layer system:

<div align="left" dir="ltr" id="bkmrk-layer-responsibility"><table><colgroup><col width="220"></col><col width="403"></col></colgroup><tbody><tr><td>### Layer

</td><td>### Responsibility

</td></tr><tr><td>Information Layer

</td><td>What data the agent can observe; what tools it may invoke at a given moment; vector storage, memory compilation, tool registries.

</td></tr><tr><td>Control Layer

</td><td>Planner / generator / evaluator roles; structured handoff artefacts; verification gates; state machines governing step sequencing.

</td></tr><tr><td>Governance Layer

</td><td>Permission boundaries, policy checks, approval gates, and the audit trail that makes autonomy observable.

</td></tr></tbody></table>

</div>