code-reviewer (CodeRabbit)
- Workflow
Hands the diff to a commercial review engine and reports what it finds: vulnerabilities, code smells, convention breaches, performance traps and likely bugs. You are buying somebody else's trained reviewer rather than running your own checklist.
A situation it fits
Your team already pays for an AI review service and likes its findings. You just want them in the terminal next to the diff instead of on a web page you have to go and open.
Skills it gets confused with
These share a family with code-reviewer (CodeRabbit), which is another way of saying they are the ones you might reach for by mistake.
- code-review (Matt Pocock)Reviews a diff since a fixed point along two axes at once: whether the code follows the repo's own documented standards, and whether it does what the originating ticket actually asked for. The second axis is the unusual one — it catches work that is clean but off-brief.
- code-reviewer (Jeff Allan)One sweep across everything at once: correctness, security, N+1 queries, naming, architecture and test coverage, returned as a single prioritised report. Breadth in one pass, where the specialised reviewers each go deep on one axis.
- security-review (Sentry)Hunts only for exploitable security holes — injection, XSS, broken auth, weak crypto — and deliberately stays quiet unless it can point at attacker-controlled input reaching a vulnerable pattern. It reports high-confidence findings only, so a clean run is not proof of safety.
Knowing what code-reviewer (CodeRabbit) does is the easy half. Telling it apart from the others under time pressure is the game.
Today's sessioncode-reviewer (CodeRabbit) is part of coderabbitai/skills. Licence: MIT. The description above was written for this game, not taken from the skill.