Skills
Place AI code review between static analysis and humans, and use test-impact analysis
The highest-leverage spot for an AI reviewer in 2026 is after linters, formatters, and static analysis but before human review: it runs on every PR in ~90 seconds, posts findings, and lets developers fix issues before a human ever sees the diff. Pair it with Test Impact Analysis so only the tests a change actually affects run, and prefer intent-based tests that re-derive when agents refactor over brittle scripts that break on every rename. Humans should never be the first line of defense against mechanical issues.
↳ Follow the thread