AI Slop Score¶
This guide explains how CloudAEye assesses the code quality of a pull request.
Overview¶
CloudAEye analyzes PR code changes for duplicate code, nonexistent APIs, dead code, missing tests, and other code quality issues when you comment @cloudaeye /slop_score in the PR timeline.
The result includes a review band and findings to help you identify changes that need closer review. The band describes code quality, not whether the code was written by AI.
Prerequisites¶
Step 1: Register¶
Sign up with CloudAEye SaaS.
Step 2: Setup Code Review¶
Setup Code Review by following the Getting Started guide.
Generate AI Slop Score¶
To analyze the code quality of your pull request, comment the following in the PR timeline:
@cloudaeye /slop_score
CloudAEye analyzes the code changes and adds a Slop Score Card to the pull request.
Slop Score Card¶
The card includes:
- A score out of 100 and its review band for the code change.
- The number of changed lines, files, and functions with behavioral changes.
- Code quality signals and their findings.
- Issue severity, details, and file locations.
Code Quality Signals¶
| Signal | Description |
|---|---|
| Duplicate code | Logic copied within the change or already available in the repository |
| Hallucinated API | Calls, imports, or environment variables that cannot be resolved |
| Dead code | Unused code or unreachable branches |
| Missing tests | Untested code, weakened assertions, or tests that restate the implementation |
| High complexity | Complexity and nesting added by the change |
| Generic naming | Names that do not describe their purpose |
| Boilerplate comments | Comments that repeat the code |
| Overengineering | Unnecessary wrappers or abstractions |
| Style inconsistency | Code that conflicts with the surrounding conventions |
Score and Review Bands¶
CloudAEye reports a score out of 100 and one of the following bands. The score is not calibrated yet, and the card says so:
- Low likelihood of slop
- Review recommended
- High likelihood of low-quality code
- Strong candidate for manual review before merge
The card identifies signals with incomplete coverage. A signal that was not measured should not be treated as having no issues.
AI Slop Score does not replace a code review. Use Review Code to check for bugs and security vulnerabilities.