<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Testing on Jack Monte — AI Engineer</title><link>https://jackmonte.com/tags/testing/</link><description>Recent content in Testing on Jack Monte — AI Engineer</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 10 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://jackmonte.com/tags/testing/index.xml" rel="self" type="application/rss+xml"/><item><title>A Green Eval Was Lying to Me</title><link>https://jackmonte.com/posts/a-green-eval-was-lying/</link><pubDate>Wed, 10 Jun 2026 00:00:00 +0000</pubDate><guid>https://jackmonte.com/posts/a-green-eval-was-lying/</guid><description>Adding strict typing to my eval harness exposed an intermittent judge failure. Fixing that exposed a second bug a passing, zero-error evaluation had been hiding the whole time. A green eval is not a correct eval.</description></item></channel></rss>