<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Rollback on Thomas Talks AI: Claude Code, AI Agents &amp; Production Engineering</title>
    <link>https://thomasdevos.com/tags/rollback/</link>
    <description>Recent content in Rollback on Thomas Talks AI: Claude Code, AI Agents &amp; Production Engineering</description>
    <generator>Hugo -- 0.145.0</generator>
    <language>en-us</language>
    <copyright>Thomas De Vos ©</copyright>
    <lastBuildDate>Fri, 01 May 2026 10:43:00 +0100</lastBuildDate>
    <atom:link href="https://thomasdevos.com/tags/rollback/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Claude Code Rollback Plans Belong in the Prompt</title>
      <link>https://thomasdevos.com/posts/2026-05-01.10.43.00-claude-code-rollback-plans-belong-in-the-prompt/</link>
      <pubDate>Fri, 01 May 2026 10:43:00 +0100</pubDate>
      <guid>https://thomasdevos.com/posts/2026-05-01.10.43.00-claude-code-rollback-plans-belong-in-the-prompt/</guid>
      <description>How to make Claude Code and AI coding agent changes safer by putting rollback plans, evidence, ownership, and human review gates into the prompt before the agent edits code.</description>
    </item>
  </channel>
</rss>
