We all want our Flutter projects to enforce their own standards, but manually writing custom lints, comprehensive tests, and maintenance scripts is a tedious chore. What if your development workflow could learn to do the heavy lifting for you?

In this lightning talk, we’ll explore how to build "almost" self-improving agentic flows tailored for the Flutter ecosystem. By integrating AI retrospective skills with automated feedback loops, we can shift from writing rigid, manual rules to creating dynamic systems that evaluate their own outputs, learn from analyzer errors, and adapt to your specific project conventions over time.