The LGTM smell is a code review smell in which reviewers submit “LGTM” (“Looks Good To Me”) reviews in place of substantive feedback — for instance when a pull request gives them too little context to understand the change, or when they are too busy to review it properly. It is counted among the practices that undermine the effectiveness of code review.
Usage
Rethinking Code Review in the Age of AI: A Vision for Agentic Code Review uses the term, drawing on earlier code review research it cites, to connect several challenges of traditional pull-request review to LGTM reviews. It links them to a number of situations: when a pull request lacks a descriptive title and description, which confuses reviewers; when missing documentation leaves reviewers unable to tell intended behavior from defects; when a missing issue link deprives them of the requirement being implemented; and when time pressure pushes them to inspect changes hastily. It also reports that prior work names reviewer availability as a direct root cause — a reviewer who is too busy with other tasks but cannot decline the review request — and that one cited study found 64.7% of pull requests across five large-scale projects reviewed without any comment, with such comment-free reviews exhibiting the LGTM smell 3.5 times more often than commented ones.
The same paper returns to the term when discussing how AI-assisted review should be evaluated, arguing that speed-centric metrics such as time-to-merge can mislead once code generation is accelerated, and that evaluation should instead account for how thoroughly review is conducted, moving beyond LGTM smells.