review

To check your code for bugs and style issues, you can run an automated code review on your files or diffs to identify correctness problems, null handling errors, and naming inconsistencies, receiving structured issue reports with severity levels and suggested fixes.

Reviewing a git diff for logic errors and security vulnerabilities involves analyzing the changed code lines to detect injection risks, hardcoded secrets, off-by-one errors, and unsafe deserialization, providing a report of critical issues with their exact file locations and recommended code snippets.

Yes, you can perform a code review on multiple programming languages at once by providing the mixed code selections, allowing the review process to uniformly evaluate tensor dtype mismatches, memory leaks, and dead code across different language syntaxes within the same structured output format.

Automated code review reports the severity and location of issues by structuring the output for every found problem into file and line numbers, a severity grade of critical, warning, or nit, a one-line description, the rationale, and a specific suggested fix.

To identify performance bottlenecks and memory leaks in your code, you execute a review pass that targets unnecessary data copies, repeated computations, and memory allocation patterns, outputting a list of flagged lines with explanations and optimized code alternatives.

If a code review finds no bugs or style issues in your script, the process explicitly states that no issues were found rather than inventing non-existent problems, ensuring you receive an accurate validation of your code's correctness and security posture.