Making Engineering Wisdom Machine-Readable: How AGENTS.md Transforms Code Review

GitHub's new AGENTS.md feature represents a paradigm shift in code review—moving beyond automated syntax checking to encode actual team judgment and architectural decisions into AI code reviewers. This matters because the most valuable aspects of code review have historically lived in human experience and institutional memory. As AI-assisted review becomes mainstream, formalizing this knowledge through well-crafted AGENTS.md files becomes critical, though the real challenge is writing specific, opinionated guidance rather than generic best practices.