<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Agent Skills on Andreas Nissen</title><link>https://andreasnissen.dev/tags/agent-skills/</link><description>Recent content in Agent Skills on Andreas Nissen</description><generator>Hugo</generator><language>en-us</language><copyright>© Andreas Nissen</copyright><lastBuildDate>Mon, 01 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://andreasnissen.dev/tags/agent-skills/index.xml" rel="self" type="application/rss+xml"/><item><title>8 Rules for Running an AI Agent Skill Library</title><link>https://andreasnissen.dev/writing/eight-rules-ai-agent-skill-library/</link><pubDate>Mon, 01 Jun 2026 00:00:00 +0000</pubDate><guid>https://andreasnissen.dev/writing/eight-rules-ai-agent-skill-library/</guid><description>&lt;p&gt;Building an agent skill is cheap. Importing one is even cheaper.&lt;/p&gt;&#10;&lt;p&gt;That is why skill libraries tend to grow faster than the discipline required to operate them. You end up with a folder full of instructions and no clear view of what half of them do, when they trigger, whether they overlap, or what they can execute.&lt;/p&gt;&#10;&lt;p&gt;After seven months running roughly 100 skills across multiple agent harnesses, the biggest shift was simple: stop treating skills as files you write. Treat them as a system you operate.&lt;/p&gt;</description></item></channel></rss>