The latest articles on DEV Community by Liedson Habacuc (@lhabacuc). https://dev.to/lhabacuc https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3259483%2F87890c8d-9c32-4068-9d97-bf2a092d221d.png https://dev.to/lhabacuc en Liedson Habacuc Wed, 18 Feb 2026 13:44:48 +0000 https://dev.to/lhabacuc/how-i-created-smartagent-and-how-it-helped-me-finally-understand-ai-agents-194h https://dev.to/lhabacuc/how-i-created-smartagent-and-how-it-helped-me-finally-understand-ai-agents-194h <p>Hi Dev.to community 👋</p> <p>I recently started diving deep into <strong>AI agents</strong> — systems that can reason, decide, and act autonomously. At first, I experimented with popular frameworks like <strong>LangChain</strong>, but I realized I didn’t fully understand <strong>what an agent actually does</strong> or how decisions and tools really work under the hood.</p> <p>So I decided to <strong>build my own library from scratch</strong>: <strong>SmartAgent</strong>.</p> <p>The goal was simple: make agents <strong>transparent, understandable, and easy to experiment with</strong>. I wanted a library that clearly separates the agent’s reasoning steps so anyone (including myself!) could actually learn how an agent works.</p> <h3> How SmartAgent works </h3> <p>SmartAgent uses a <strong>three-phase flow</strong>:</p> <ul> <li> <strong>Analysis</strong> – the agent reasons about the problem</li> <li> <strong>Execution</strong> – the agent decides what to do and uses tools</li> <li> <strong>Response</strong> – the agent produces a final answer</li> </ul> <p>By designing it this way, I could finally understand:</p> <ul> <li>What an agent <em>really is</em> (not just prompts)</li> <li>How decisions are made step by step</li> <li>How tools fit into the reasoning loop</li> <li>How to design my own agents intentionally</li> </ul> <h3> Why I’m sharing it </h3> <p>I built SmartAgent not just for myself, but as a <strong>learning tool for others</strong>. If you’re curious about AI agents, want to experiment, or even contribute, it’s approachable and educational.</p> <p>It’s also open-source, so anyone can explore, tweak, or add new tools.</p> <p><strong>Try it out:</strong></p> <ul> <li>GitHub: <a href="https://github.com/lhabacuc/smartagent" rel="noopener noreferrer">https://github.com/lhabacuc/smartagent</a> </li> <li>PyPI: <a href="https://pypi.org/project/smartagent-sf/" rel="noopener noreferrer">https://pypi.org/project/smartagent-sf/</a> </li> </ul> <p>I’d love to hear how other people are experimenting with agents, and any feedback or ideas are very welcome!</p> ai python opensource agents