How do you prevent AI-generated code from drifting away from your conventions over time?

A developer building Kumiko, an opinionated SaaS framework, describes the challenge of AI-generated code drifting from project conventions over time. While initial AI-generated features match codebase patterns, subsequent features accumulate subtle inconsistencies that linting and code review fail to fully prevent. The team is exploring framework constraints as a partial solution but acknowledges the problem remains unsolved.

We've been generating production features with AI for a while now — auth flows, billing hooks, notification handlers. And we've hit a pattern we don't have a good answer to yet. The first feature the AI generates looks great. It reads the codebase, picks up the patterns, and the output looks like something a senior dev wrote. The tenth feature? Less so. Small inconsistencies creep in: None of it is wrong. All of it is subtly inconsistent. AGENTS.md / CLAUDE.md — helps, but gets stale and doesn't scale with the codebase Code review — catches it, but defeats some of the speed advantage Linting + formatting — catches easy stuff, misses semantic drift What we haven't solved: giving the AI a "living" representation of your conventions that stays current as the codebase evolves. We're building Kumiko — an opinionated SaaS framework — partly as an answer to this. If the framework constrains what's possible, drift has less surface area. But I'm not convinced that fully solves it either. What's your approach?