https://blog.yushi91.com/blog/context-management/Recent content in Context-Management on Yushi's BlogHugoen-USCopyright © 2025, Yushi Cui.Tue, 05 May 2026 10:00:00 +1200https://blog.yushi91.com/blog/three-stages-agent-workflow/Tue, 05 May 2026 10:00:00 +1200https://blog.yushi91.com/blog/three-stages-agent-workflow/<p><img src="https://blog.yushi91.com/agent-workflow-three-stages.webp" alt="Three stages of an agent workflow" loading="lazy"> </p> <h2 id="the-mental-model">The mental model</h2> <p>I want to talk about the way I currently think about agent workflows. Three stages. Each one solves a different problem, and you need all three before you get something that can run for hours without falling over.</p> <p>The three stages are prompt engineering, context management, and the harness. I&rsquo;ll go through each one in the order I learned them.</p> <h2 id="stage-one-prompt-engineering">Stage one: prompt engineering</h2> <p>The goal here is simple. Give the agent a clear instruction so it knows what to do.</p>