Skip to main content
| On GitHub repositories CodeRabbit analyzes pull requests for signals of low-quality, AI-generated content. If the PR is classified as slop, CodeRabbit notes this in the PR Walkthrough comment. You can also configure a label in your settings to get the suspicious PRs labeled automatically. Analysis runs only on the first full, non-incremental review of a pull request, and only when reviews.slop_detection.enabled is on. No additional setup is required beyond enabling the feature.

Configuration

Configure the slop_detection section in your .coderabbit.yaml file:
.coderabbit.yaml

Disabling Slop Detection

To disable Slop Detection on a repository, set enabled to false or disable it in the web UI:
.coderabbit.yaml

Label flagged PRs

By default, Slop Detection does not apply any label, it only adds a note to the PR Walkthrough comment. To also apply a label to flagged PRs, set the label field:
.coderabbit.yaml

Frequently asked questions

Yes. Slop Detection runs on private repositories for organizations on the Essentials plan and above. Public repositories are eligible across all plans.
Yes. Service or automation accounts are exempt. Authors are also exempt when their repository relationship is collaborator, contributor, member, or owner.
No. Slop Detection does not block merges. When a PR is classified as slop, a note is added to the PR Walkthrough comment. If a label is configured, it is also applied so maintainers can filter and triage PRs at their discretion.
Not at the moment. CodeRabbit is currently gathering data and tuning the detection mechanism to ensure accuracy.

What’s next

Pre-Merge Checks

Enforce quality gates and custom requirements before pull requests are merged

Automatic review controls

Fine-tune which PRs CodeRabbit reviews automatically using labels, branches, and author filters

Configuration reference

See the full slop_detection schema with types, defaults, and constraints