<?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>AI Penetration Test on Cybersecurity blog by Gabor Szekeres</title><link>https://gaborszekeres.com/categories/ai-penetration-test/</link><description>Recent content in AI Penetration Test on Cybersecurity blog by Gabor Szekeres</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Tue, 15 Sep 2026 13:11:40 +0200</lastBuildDate><atom:link href="https://gaborszekeres.com/categories/ai-penetration-test/index.xml" rel="self" type="application/rss+xml"/><item><title>What Sentinel caught when I attacked my own AI app</title><link>https://gaborszekeres.com/blog/what-sentinel-caught-when-i-attacked-my-own-ai-app/</link><pubDate>Tue, 15 Sep 2026 13:11:40 +0200</pubDate><guid>https://gaborszekeres.com/blog/what-sentinel-caught-when-i-attacked-my-own-ai-app/</guid><description>Half of the instruction-override attempts were never flagged. Here&amp;#39;s what that means for AI detection strategy.</description></item></channel></rss>