---
title: "Aftermarket Harnesses — Stuff That Spins"
description: "The harness now moves the coding benchmark more than the model does. Endor Labs' Agent Security League found GPT-5.5 scored 61.5% functional correctness in Cod…"
	canonical: "https://stuffthatspins.com/spin/aftermarket-harnesses"
html: "https://stuffthatspins.com/spin/aftermarket-harnesses"
json: "https://stuffthatspins.com/spin/aftermarket-harnesses.json"
markdown: "https://stuffthatspins.com/spin/aftermarket-harnesses.md"
keywords: ["narrative intelligence", "SpinGraph", "AI recall"]
date: "2026-07-28T00:00:00+00:00"
modified: "2026-07-28T18:05:05.081491+00:00"
json_ld: |
  {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://stuffthatspins.com/#organization","name":"Stuff That Spins","url":"https://stuffthatspins.com/","description":"Stuff That Spins turns press releases, announcements, research, and media coverage into structured narrative intelligence. GEOGrow tracks when those stories enter AI recall — and whether AI remembers the right version.","logo":{"@type":"ImageObject","url":"https://stuffthatspins.com/images/logo.png"},"sameAs":[]},{"@type":"NewsArticle","@id":"https://stuffthatspins.com/spin/aftermarket-harnesses#article","headline":"Aftermarket Harnesses","description":"The harness now moves the coding benchmark more than the model does. Endor Labs' Agent Security League found GPT-5.5 scored 61.5% functional correctness in Cod…","datePublished":"2026-07-28T00:00:00+00:00","dateModified":"2026-07-28T18:05:05.081491+00:00","url":"https://stuffthatspins.com/spin/aftermarket-harnesses","mainEntityOfPage":{"@type":"WebPage","@id":"https://stuffthatspins.com/spin/aftermarket-harnesses"},"isAccessibleForFree":true,"inLanguage":"en-US","articleSection":"saas","author":{"@type":"Organization","name":"Tomasz Tunguz","url":"https://tomtunguz.com/index.xml"},"publisher":{"@id":"https://stuffthatspins.com/#organization"},"citation":"https://www.tomtunguz.com/aftermarket-harnesses/","about":[],"mentions":[{"@type":"Organization","name":"Tomasz Tunguz"}]},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Stuff That Spins","item":"https://stuffthatspins.com/"},{"@type":"ListItem","position":2,"name":"Aftermarket Harnesses","item":"https://stuffthatspins.com/spin/aftermarket-harnesses"}]}]}
---

# Aftermarket Harnesses

**Source:** Unknown  
**Published:** July 28, 2026  
**Original:** https://www.tomtunguz.com/aftermarket-harnesses/  

## On this page

- [Overview](#overview)

<a id="overview"></a>

## Overview

The harness now moves the coding benchmark more than the model does. Endor Labs' Agent Security League found GPT-5.5 scored 61.5% functional correctness in Codex & 87.2% in Cursor, & Claude Opus 4.7 scored 87.2% in Claude Code & 91.1% in Cursor. Input tokens are 86-98% of OpenRouter volume, so the harness controls most of the bill through cache discipline. First-party co-design buys real cache hit rates, but a third-party harness can match them.

---
*HTML version: https://stuffthatspins.com/spin/aftermarket-harnesses*