From GitHub · · 1m
Harness Engineering 101
Building AI agents from first principles with a simple harness model.
In brief
Building AI agents from first principles with a simple harness model. Harness engineering creates AI agents by patching a stateless JSON array loop with minimal, concrete fixes. Originally reported by GitHub.
The LLM is the brain
The LLM is the brain. The harness is the body.
Version 1.0, September 2026
Version 1.0, September 2026. A blog series about building AI agents from first principles, published as a book.
Stateless LLM API
An LLM API is stateless. Every turn, you send the entire conversation as a JSON array and get text back.
The Harness
A harness is the program that builds, maintains, and protects that array. Everything the field calls 'agents' is a set of patches to that one loop.
16 chapters
This series develops that one idea across 16 chapters, growing a ~300-line toy agent in plain Python as it goes.
One Code
The patterns here are developed in full in One Code, an open-source coding agent built on pi.
What it means
Harness engineering creates AI agents by patching a stateless JSON array loop with minimal, concrete fixes.
