code-review-checklist

To review code for security vulnerabilities and best practices, use a structured checklist covering correctness, performance, testing, and documentation, while identifying common anti-patterns and leaving standardized feedback.

To check code for prompt injection risks before a pull request, evaluate the LLM logic against a structured code review checklist that specifically addresses prompt injection, output sanitization, and hallucination mitigation.

Common anti-patterns to look for during a code review include code smells and structural issues that violate best practices in correctness, security, performance, code quality, testing, and documentation, as outlined by a standardized checklist.

This code review checklist covers both general software engineering and AI-specific code, providing baseline checks for correctness, security, and performance, alongside specialized guidelines for LLM prompt injection and hallucination risks.

To write effective review comments for a teammate's code, follow a standardized comment guide that identifies common anti-patterns and addresses correctness, security, and performance issues clearly and constructively.