Three PRs to Rowboat in 24 Hours: Lessons in AI-Assisted Contribution

A developer recently shared their experience using AI coding agents to contribute three pull requests to the open-source project Rowboat within a 24-hour period. The contributions addressed specific UI bugs, server stability issues, and the addition of a new CLI feature. The author highlights a critical workflow: using agents to generate code while maintaining rigorous human oversight to catch errors, such as incorrect branching. Beyond the technical implementation, the author emphasizes the importance of auditing existing issues rather than relying solely on labels, noting that many open issues in popular repositories are often abandoned or already resolved. The post serves as a practical guide for contributors, advising them to act as 'hostile senior engineers' when reviewing AI-generated diffs and to verify the actual state of the codebase before attempting to claim tasks, ensuring that contributions remain meaningful and clean.