# Agent Harness

> An agent harness is the system around the model that makes it an agent — the loop, tools, context management, permissions, and recovery machinery.

**An agent harness is everything around the model that turns it into a working agent: the execution loop, tool definitions, context management, permissions, error handling, and recovery — the machinery that converts model decisions into safe, observed actions.**

The term sharpened as the industry learned that **model quality and agent quality are different axes**. A harness determines what the model sees each turn (context assembly, [compaction](/guides/configuration/claude-code-memory-context), memory), what it can do ([tools](/guides/concepts/production-tool-calling) and their schemas), what it may do ([permissions](/guides/configuration/claude-code-settings-permissions) and gates), and how failures feed back (errors as observations, retries, loop detection). Identical models in different harnesses produce visibly different agents — which is why coding-agent comparisons increasingly evaluate model+harness pairs, and why [Claude Code](/guides/getting-started/what-is-claude-code)'s edge is co-tuning both sides.

The word now anchors real decisions: *adopt a harness* (Claude Code, OpenCode, Letta Code — [the comparison axis](/guides/comparisons/claude-code-vs-opencode)), *build on one* (the [Claude Agent SDK](/guides/advanced/claude-agent-sdk-tutorial) is "the harness as a library"), or *assemble your own* from [frameworks](/guides/concepts/agent-frameworks-2026). And it names the discipline around it: [agent engineering](/glossary/agent-engineering) is, in one phrase, harness engineering.

---

_Source: https://agentscamp.com/glossary/agent-harness — Term on AgentsCamp._