<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Instructionsplitting on FindPicked</title><link>https://findpicked.com/tags/instructionsplitting/</link><description>Recent content in Instructionsplitting on FindPicked</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sat, 15 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://findpicked.com/tags/instructionsplitting/index.xml" rel="self" type="application/rss+xml"/><item><title>Defending AI Agents from MCP Instruction-Splitting Attacks</title><link>https://findpicked.com/blog/mcp-instruction-splitting-attack-defense/</link><pubDate>Sat, 15 Aug 2026 00:00:00 +0000</pubDate><guid>https://findpicked.com/blog/mcp-instruction-splitting-attack-defense/</guid><description>&lt;p&gt;The increasing sophistication of &lt;strong&gt;AI agents&lt;/strong&gt; interacting with external systems via protocols like the &lt;strong&gt;Model Context Protocol (MCP)&lt;/strong&gt; has unfortunately created novel attack vectors. Instruction-splitting attacks, originating from malicious &lt;strong&gt;MCP&lt;/strong&gt; servers, represent a critical security threat, capable of compelling &lt;strong&gt;AI coding agents&lt;/strong&gt; to exfiltrate sensitive data. This article will detail the technical mechanisms of these attacks and provide practical, actionable defense strategies for developers to safeguard their &lt;strong&gt;AI coding agents&lt;/strong&gt;.&lt;/p&gt;</description></item></channel></rss>