I used to enforce heavy pre-commit hooks to optimise for fast feedback. These days, I push most of those checks to CI.
This post isn’t about tooling so much as it is about flow state, cognitive load, and learning the cost of enforcing the right checks at the wrong time.
Coding agents don’t fail because the model is weak. They fail because the environment they land in is noisy, inconsistent or ambiguous. This post shows senior engineers and tech leads how to reshape one slice of a real-world repo so agents behave predictably instead of drifting.
On a long-running headless multisite project, our code review process slowly spiralled out of control, with 100+ open PRs, slow turnarounds, constant context-switching, and frustrated engineers. We fixed it by reframing how we think about reviews, creating a clear hierarchy, and making readability the foundation. This is the story of how we made code reviews calmer, faster, and far more...