Qodo Solutions Engineer Cole Gannaway: Using Software Graphs to Detect Cross-Repo Blast Radius in AI Code Review
In this AI Engineer podcast episode, Qodo solutions engineer Cole Gannaway discusses how the bottleneck in AI-assisted development has shifted from code generation to review. He explains how a one-line data model change can break multiple client types (web, mobile, CLI, MCP) and demonstrates how building a software graph of repositories reveals this hidden 'blast radius.' The episode also covers turning tribal knowledge and PR comments into custom review rules and why the agent that wrote the code should not grade it.
AI 代码生成速度的提升已将工程瓶颈从编写转移至审查。在最新一期 AI Engineer 播客中，Qodo 解决方案工程师 Cole Gannaway 探讨了这一转变带来的具体挑战。他指出，即使是一行数据模型的微小改动，若未充分理解其跨仓库依赖关系，也可能悄然破坏 Web、移动端、CLI 和 MCP 客户端。
Gannaway 提出了一种通过构建软件图谱（software graph）来可视化这些隐藏依赖的方法，从而揭示代码变更的“爆炸半径”。此外，他还讨论了如何将团队内部的隐性知识（tribal knowledge）和过往 PR 评论转化为自定义审查规则，以及为何不应让生成代码的代理自行评估其质量。对于正在应对 AI 辅助开发规模化挑战的工程团队，该集提供了关于上下文管理和自动化审查的具体实践视角。