code-review

Review code changes for defects, security risks, and boundary violations.

4|Updated May 16, 2026
One-click install
npx skills add https://github.com/machenjie/rd-skills --skill code-review-machenjie
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/machenjie/rd-skills/tree/main/src/foundation/capabilities/code-review
Command: npx skills add https://github.com/machenjie/rd-skills --skill code-review-machenjie

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps teams prevent production-impacting defects by grounding every review decision in evidence, test behavior, or reproducible reasoning.

Core Features & Use Cases

  • Evidence-grounded defect detection: Identify correctness, maintainability, boundary violations, security risks, and hallucinated APIs with explicit file evidence or named failure modes.
  • Severity classification with merge policy: Classify findings as Critical/High/Medium/Low and enforce blocking rules for security and user-impacting failures.
  • Test and error-handling verification: Ensure material behavior changes have adequate test coverage and that failures are handled explicitly without misleading success or swallowed errors.
  • Security surface coverage: Review authentication/authorization, injection vectors, data exposure, crypto misuse, and error/stack trace handling—then document non-findings when applicable.

Quick Start

Use code-review to assess a specific pull-request diff or code change by checking correctness, security surfaces, boundary rules, hallucinated APIs, and whether tests cover changed behavior.

Frequently Asked Questions about code-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check a pull request diff for security risks and architectural boundary violations?

To check a pull request diff for security risks and architectural boundary violations, review the code changes using evidence-backed reasoning and scenario-based checks to identify material defects and enforce merge policies.

What is the best way to detect hallucinated APIs and dependency risks in generated code?

Detecting hallucinated APIs and dependency risks in generated code requires verifying the existence of APIs, configurations, and libraries, grounding every review decision in explicit file evidence or named failure modes.

How do I enforce test coverage and error handling validation for changed behavior in a refactor?

Enforce test coverage and error handling validation for refactored behavior by ensuring material changes have adequate tests and that failures are handled explicitly without misleading success or swallowed errors.

Does this code review approach classify defect severity for production changes?

Yes, this code review approach classifies defect severity for production changes by categorizing findings as Critical, High, Medium, or Low and enforcing blocking rules for security and user-impacting failures.

What security surfaces should I review when upgrading dependencies or modifying production code?

When upgrading dependencies or modifying production code, review security surfaces including authentication, authorization, injection vectors, data exposure, crypto misuse, and error stack trace handling, documenting non-findings when applicable.

When should I not use automated evidence-grounded reasoning for code reviews?

Automated evidence-grounded reasoning for code reviews is not suitable when you lack explicit merge policy expectations or when evaluating changes that do not involve material behavior, security surfaces, or architectural boundaries.