Why skills are the best yet most dangerous things in AI right now

<p>A well-behaved AI agent deleted a company&#39;s entire production database, and every backup with it, in nine seconds. Nobody hacked it. It was doing exactly what it had been told, with a key it should never have been holding. </p><p><br></p><p>And the thing that made it possible is now one command away from everyone in your company, in marketing, sales and support, not just engineering. In under a year, installable AI agent skills have gone from a niche developer trick to an open industry standard.</p><p><br></p><p>In this episode of In The Loop, I&#39;m looking at why the exact thing that makes skills brilliant is the thing that quietly turns them into a control problem. I get into why this is shadow IT all over again except the shadow can now act, and why the fix isn&#39;t to ban any of it.</p><p>⏭️ Episode highlights</p><p>(00:45) – The database gone in nine seconds</p><p>(02:10) – One line that skips engineering</p><p>(04:00) – Skills vs MCP, what&#39;s actually dangerous</p><p>(08:00) – Shadow IT is back, but it can act now</p><p>(09:45) – PocketOS and Replit: the receipts</p><p>(11:30) – The middle path that actually fixes it</p><p>(12:40) – Where this leaves you</p>