91% of developers use AI tools. Your repo is accumulating technical debt RIGHT NOW.

PR Reviews

Quality Radar

Quality Radar provides a multi-dimensional view of code quality, scoring each PR across five key dimensions.

The Five Dimensions

Each dimension is scored from 0 to 100 based on the findings in your PR:

Security

95/100

Evaluates authentication, authorization, input validation, secret handling, and known vulnerability patterns.

Performance

88/100

Checks for N+1 queries, missing indexes, unbounded queries, memory leaks, and blocking operations.

Type Safety

92/100

Assesses type coverage, proper typing of function parameters and returns, and avoidance of type escape hatches.

Readability

85/100

Reviews naming conventions, code structure, documentation, and consistency with your codebase patterns.

Test Quality

90/100

Evaluates test coverage for new code, assertion quality, test isolation, and edge case handling.

Score Interpretation

95–100Excellent — production-ready code
85–94Good — minor improvements recommended
70–84Needs attention — significant issues found
Below 70Critical — blocking issues must be addressed

Codebase Reuse Metric

In addition to the five dimensions, the Quality Radar tracks a codebase reuse metric — how well new code leverages existing utilities, patterns, and abstractions in your repository. Low reuse scores often indicate framework reinvention, a hallmark of AI-generated code.

Quality Radar in Enterprise

On Hyperdrive and Interstellar plans, Quality Radar scores are tracked over time, giving teams insight into code quality trends across the organization. Use the Analytics dashboard to monitor dimension scores by repository, team, or time period.