I built Bubo because I’m tired of AI reviewers flooding PRs with false positives and repeating them after maintainers explain why they’re wrong. Bubo posts evidence-backed findings or LGTM and learns from repository feedback so noise drops over time. It supports GitHub and GitLab.