What an AI agent harness does, in one short TypeScript loop

Description and use

<p>What is an AI harness? It might be the most used and least explained term in AI right now.</p> <p>Here's my short answer. The model writes text. The agent harness turns that text into actions.</p> <p>A model can't open a file, run a command or click a button by itself. It generates output, and part of that output can ask for a tool. So when Claude Code fixes a failing test, the model never touches your code directly. It asks for the tests to run, and the harness runs them. The model's choices still change your files.</p> <p><a href="https://argbe.tech/blog/ai-agent-harness?utm_source=devto&

Problem

<p>What is an AI harness? It might be the most used and least explained term in AI right now.</p> <p>Here's my short answer. The model writes text. The agent harness turns that text into actions.</p> <p>A model can't open a file, run a command or click a button by itself. It generates output, and part of that output can ask for a tool. So when Claude Code fixes a failing test, the model never touches your code directly. It asks for the tests to run, and the harness runs them. The model's choices still change your files.</p> <p><a href="https://argbe.tech/blog/ai-agent-harness?utm_source=devto&

Outcome

<p>What is an AI harness? It might be the most used and least explained term in AI right now.</p> <p>Here's my short answer. The model writes text. The agent harness turns that text into actions.</p> <p>A model can't open a file, run a command or click a button by itself. It generates output, and part of that output can ask for a tool. So when Claude Code fixes a failing test, the model never touches your code directly. It asks for the tests to run, and the harness runs them. The model's choices still change your files.</p> <p><a href="https://argbe.tech/blog/ai-agent-harness?utm_source=devto&

Evidence

What an AI agent harness does, in one short TypeScript loop

Source