<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet href="/rss/styles.xsl" type="text/xsl"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>REDHEXX Research</title><description>Research notes on the MCP and agent attack surface from REDHEXX.</description><link>https://redhexx.com/</link><language>en-us</language><atom:link href="https://redhexx.com/rss.xml" rel="self" type="application/rss+xml"/><item><title>Mapping the MCP attack surface: what we look for</title><link>https://redhexx.com/research/mapping-the-mcp-attack-surface/</link><guid isPermaLink="true">https://redhexx.com/research/mapping-the-mcp-attack-surface/</guid><description>A first note on the exposure classes an AI-native security review should cover, and why the tools your agents can reach are the part most reviews still skip.</description><pubDate>Sun, 05 Jul 2026 00:00:00 GMT</pubDate></item></channel></rss>