<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Coding Agents on CybersecurityOS</title><link>http://www.cybersecurityos.net/tags/coding-agents/</link><description>Recent content in Coding Agents on CybersecurityOS</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 28 Oct 2026 00:30:38 +0000</lastBuildDate><atom:link href="http://www.cybersecurityos.net/tags/coding-agents/index.xml" rel="self" type="application/rss+xml"/><item><title>The Harness Is the Product: What Cloudflare's Vulnerability Harness Teaches AppSec Teams</title><link>http://www.cybersecurityos.net/posts/ai-devsecops/cloudflare-vulnerability-harness-security-audit-skill/</link><pubDate>Wed, 28 Oct 2026 00:00:00 +0000</pubDate><guid>http://www.cybersecurityos.net/posts/ai-devsecops/cloudflare-vulnerability-harness-security-audit-skill/</guid><description>&lt;p&gt;Most of the conversation about AI and vulnerability research is about models. Which one found the most bugs, which one is gated, which one just topped a benchmark. Cloudflare&amp;rsquo;s security team spent a long post on June 18, 2026 arguing that&amp;rsquo;s the wrong thing to focus on. In &lt;a href="https://blog.cloudflare.com/build-your-own-vulnerability-harness/"&gt;Build your own vulnerability harness&lt;/a&gt;, the point is simple: &amp;ldquo;Our approach centers on the harness, not the model.&amp;rdquo;&lt;/p&gt;
&lt;p&gt;Then they did something more useful than writing about it. They open-sourced the skill the harness grew out of as &lt;a href="https://github.com/cloudflare/security-audit-skill"&gt;cloudflare/security-audit-skill&lt;/a&gt;, MIT-licensed, installable in one command. This post covers both: how the harness is structured, what you actually get in the repo, and how I&amp;rsquo;d put it to work.&lt;/p&gt;</description></item></channel></rss>