skills.rest
voidful
voidfulCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

code-review

Reviews code diffs for correctness bugs, regressions, missing tests, and risky assumptions.

1|Updated Apr 8, 2026
Search Tags:#testing#code review#verification#quality#regression#pre-commit#correctness
One-click install
npx skills add https://github.com/voidful/Aixlarity --skill code-review-voidful
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/voidful/Aixlarity/tree/main/.aixlarity/skills/code-review
Command: npx skills add https://github.com/voidful/Aixlarity --skill code-review-voidful

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Code authors often miss bugs in their own work, and self-verification rarely catches correctness issues, regressions, or untested paths. This Skill provides a fresh-perspective review process that systematically inspects changes before they are committed. ## Core Features & Use Cases - Prioritized Findings: Ranks issues by severity (CRITICAL, WARNING, INFO) across correctness bugs, behavioral regressions, missing tests, risky assumptions, and security concerns. - Structured Output: Produces actionable findings with file path, line number, risk explanation, and suggested fixes. - Use Case: After an AI agent or developer implements a feature touching multiple files, trigger a review before committing to catch unchecked errors, hardcoded values, or missing test coverage. ## Quick Start Review the changes I just made and flag any correctness bugs, regressions, or missing tests before I commit.

Frequently Asked Questions about code-review

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

FAQPage Schema
How do I review code changes before committing?▼

Trigger a code review after implementing a feature or fix by asking to review the diff or changed files. The review checks correctness bugs, regressions, missing tests, and risky assumptions, then reports findings ranked by severity with file locations and suggested fixes.

What does an automated code review check for?▼

It checks five priorities in order: correctness bugs, behavioral regressions, missing tests, risky assumptions like hardcoded values or race conditions, and security issues such as prompt injection or credential exposure. Style issues handled by linters are explicitly excluded.

When should I request a code review instead of implementation?▼

Request a review after completing a feature or bug fix, especially when the task touched two or more files. It is also appropriate whenever you ask to check, look at, or find what is wrong with existing code rather than modify it.

What are the limitations of an AI code review?▼

The review does not rewrite code or fix issues itself; it only reports findings with suggested fixes. It also intentionally skips style nitpicks that linters should catch, so it complements rather than replaces linting and test suites.

How are code review findings formatted?▼

Each finding includes a severity level (CRITICAL, WARNING, or INFO), a description, the file path and line number, an explanation of the risk, and a suggested fix when one is obvious. The summary stays short and focuses on actionable items.

Related Skills

Explore More Software Engineering Skills→
sizukutamagosizukutamago

tdd

Apply test-driven development cycles with requirements.md and test-runner agents.

Community
Advanced
fzymgc-housefzymgc-house

address-findings

Apply code review fixes and manage dependencies in Git or Jujutsu repositories.

Official
Advanced
koralliskorallis
90

debugging-systematic

Diagnose bugs using a four-phase root cause analysis framework.

Community
Advanced
in-the-loop-labsin-the-loop-labs
48

local

Open uncommitted local code changes in the pair-review web UI.

Official
Basic