Code Reviewer FAQs

Question 1

How does the Code Reviewer handle subjective 'nitpicks'?

Accepted Answer

The skill uses a strict confidence scoring system. Only issues with a certainty score of 80% or higher are reported, ensuring that feedback is objective and impactful rather than speculative.

Question 2

How does it decide whether to approve or request changes?

Accepted Answer

Approval is granted when all verification checks pass and business logic is sound. Changes are requested for critical issues like security holes, correctness bugs, or missing tests.

Question 3

What dimensions of quality does the review cover?

Accepted Answer

It evaluates code across six specific dimensions: Correctness, Safety (Security), Maintainability, Testability, Performance, and Standards Compliance.

Question 4

Does this skill report pre-existing issues in the codebase?

Accepted Answer

No, the skill includes a critical filtering rule to ignore pre-existing issues not introduced by the current change, focusing specifically on the impact of the new code.

Question 5

Can I use this instead of a linter or automated test suite?

Accepted Answer

No, this skill is designed to be used after automated verification. It provides human-like reasoning for architectural and logic issues that automated tools often miss.

Code Reviewer

Key Features

Use Cases

Code Reviewer

Key Features

Use Cases