<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Security Engineering on</title><link>https://www.ngallo.it/tags/security-engineering/</link><description>Recent content in Security Engineering on</description><generator>Hugo</generator><language>en-GB</language><managingEditor>nicola.gallo@nitroagility.com (Nicola Gallo)</managingEditor><webMaster>nicola.gallo@nitroagility.com (Nicola Gallo)</webMaster><lastBuildDate>Tue, 17 Nov 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://www.ngallo.it/tags/security-engineering/index.xml" rel="self" type="application/rss+xml"/><item><title>Beyond Permissions: Verifiable Execution for AI Agents and Distributed Systems</title><link>https://www.ngallo.it/blog/2026-11-17/verifiable-execution-integrity-pic-coss/</link><pubDate>Tue, 17 Nov 2026 00:00:00 +0000</pubDate><author>nicola.gallo@nitroagility.com (Nicola Gallo)</author><guid>https://www.ngallo.it/blog/2026-11-17/verifiable-execution-integrity-pic-coss/</guid><description>&lt;figure class="post-banner"&gt;
 &lt;img src="https://www.ngallo.it/images/2026-11-17/execution-plane-integrity.png"
 alt="A valid permission check does not prove a valid data path: verifiable execution for AI agents and distributed systems."
 loading="lazy"&gt;
&lt;/figure&gt;
&lt;p&gt;&lt;strong&gt;Authorization answers what an executor may do. Execution integrity asks whether the result really came from the permitted execution.&lt;/strong&gt;&lt;/p&gt;
&lt;p&gt;Imagine an AI agent that must call Tool A and pass the result to Tool B. If access to A is denied, the security layer has done its job. Yet the agent can still produce a plausible answer and send it to B &lt;em&gt;as though A had executed&lt;/em&gt;. Even when A is available, the agent may change an inconvenient result before forwarding it.&lt;/p&gt;</description></item></channel></rss>