<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>Non-Human Identity on victorh.dev</title>
		<link>/tags/non-human-identity/</link>
		<description>Recent content in Non-Human Identity on victorh.dev</description>
		<generator>Hugo</generator>
		<language>en</language>
		
		
		
			<copyright>Copyright © {year} Victor Hernandez Guzman. All rights reserved. Created with :heart: and :coffee:. </copyright>
		
		
			<lastBuildDate>Mon, 26 Oct 2026 01:05:03 -0700</lastBuildDate>
		
			<atom:link href="/tags/non-human-identity/index.xml" rel="self" type="application/rss+xml" />
			<item>
				<title>From Non-Human Identity to Agent Identity: Where the Threshold Sits</title>
				<link>/blog/nhi-to-agentidentity/</link>
				<pubDate>Mon, 26 Oct 2026 01:05:03 -0700</pubDate>
				<guid>/blog/nhi-to-agentidentity/</guid>
				<description>Enterprise IAM treats machine identities as predictable workloads, but AI agents choose their own actions at runtime. This article offers a three-part threshold test: model-directed control flow, runtime capability binding, and delegated human authority across hops. It describes the four failures that follow when agents run on service accounts, and shows how a task-bound, short-lived composite token fixes them.</description>
			</item>
	</channel>
</rss>
