Two AI code reviewers disagree sharply on the same Fastify PR

On October 3, 2026, DEV Community published an account by a developer who ran two AI code reviewers from different vendors over the same merged Fastify pull request and got sharply different results.

The pull request is fastify#6716, "feat: add support of onMaxParamLength", which the author says was merged in May. According to the post, one of the two reviewers found nothing at all. The second one surfaced a behavior change that, the author writes, is still in Fastify today. That same reviewer, however, was wrong about one important detail.

The post says it presents what happened, how the author checked the finding, and what he changed in his own tool because of it. Those verification steps and the tool changes are part of the write-up rather than something summarized up front.

The account does not name the two vendors behind the reviewers. It also does not spell out the behavior change itself or identify the detail the second reviewer got wrong, at least not in the portion quoted. So the story rests, as of this report, on one developer's test of a single merged PR: the same input sent to two vendor-supplied reviewers, one silent result, and one finding that is both still live in Fastify and partly incorrect. No vendor response and no independent confirmation of the finding appear in the material.