Code Quality Reviewer

Problem

Code review often stays at the surface level: naming, indentation, and formatting are checked, while the real risks are missed. These include null handling, out-of-bounds access, race conditions, hardcoded secrets, inefficient for loops, and repeated queries. Engineers need a repeatable review framework, not another vague prompt that produces generic feedback.

How It Works

The skill evaluates a code snippet across four fixed dimensions:

• Style and readability: naming, structure, comments, and function length.
• Potential bugs: edge cases, null / undefined, array indexes, recursion, and concurrency safety.
• Performance: algorithmic complexity, resource use, and optimization targets.
• Security: SQL injection, XSS, authorization checks, and hardcoded API keys.

It returns a Markdown report with an executive summary, detailed findings, and a conclusion. Warnings, bugs, and security issues are marked with symbols, and each finding includes a concrete suggestion.

Boundaries

It is useful for JavaScript, TypeScript, Python, Java, and Go snippets. When context is missing, such as API contracts, database models, caller constraints, or business invariants, conclusions are inferences based on visible code only. For high-risk areas like encryption, payments, or authorization, treat the output as a review checklist, not a final security audit.