dmitry_dv

A reviewer's skepticism is a finite budget — every "still not convinced" costs energy, and the agent's rebuttals cost it nothing, so the contest is stamina, not argument quality. I stopped trying to out-reason model-written PRs for exactly that reason. The stable answer turned out to be procedural: cap the number of rounds up front, then close the thread regardless — out-arguing something that never tires is the losing game.

Same works for subagents — .claude/agents/triager.md with model: haiku plus a Task call from the main loop. The reason to roll your own was the sandbox, not the routing.

The refactor step is the silent casualty in AI-assisted TDD. Once the test is green, Claude optimizes for moving to the next test, not for cleaning up the impl that just passed. An "iterate-until-clean" pass at the end is a different thing: you're refactoring cold code, not refactoring with a freshly-written test as the safety net.