code-review

Reviews code changes through parallel code-reviewer and architect agent lanes with severity-rated findings.

Updated Feb 1, 2024
One-click install
npx skills add https://github.com/ogiboy/portfolio-app --skill code-review-ogiboy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/ogiboy/portfolio-app/tree/main/.codex/skills/code-review
Command: npx skills add https://github.com/ogiboy/portfolio-app --skill code-review-ogiboy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reviewing pull requests and code changes manually is slow and inconsistent, and single-perspective reviews often miss security vulnerabilities or architectural tradeoffs. This Skill runs a structured two-lane review that produces severity-rated findings and a deterministic merge-readiness verdict. ## Core Features & Use Cases - Parallel Review Lanes: Delegates to a code-reviewer agent for security, quality, performance, and maintainability checks, plus an architect agent for design-tradeoff analysis. - Severity-Rated Findings: Classifies issues as CRITICAL, HIGH, MEDIUM, or LOW with file:line locations and concrete fix recommendations. - Deterministic Merge Gating: Combines lane outputs into APPROVE, REQUEST CHANGES, or COMMENT verdicts using explicit rules, with architect BLOCK status forcing a change request. - Use Case: Before merging a pull request, run the review to get a synthesized report covering both code-level issues and architectural watch items, with approval blocked if independent review evidence is unavailable. ## Quick Start Ask the assistant to review the current git diff or a specific set of changed files before merging your pull request.

Frequently Asked Questions about code-review

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

FAQPage Schema
How do I run a code review on my pull request changes?

Request a review of your git diff or specific changed files, and the skill launches parallel code-reviewer and architect agent lanes. It returns a report with issues grouped by severity, file:line locations, fix suggestions, and a final merge recommendation.

What does an automated code review check for?

The review checks security issues like hardcoded secrets, injection, XSS, and CSRF, plus code quality, complexity, duplication, performance problems like N+1 queries, and maintainability concerns. A separate architect lane evaluates boundaries, coupling, and design tradeoffs.

How are code review severity levels decided?

Findings are rated CRITICAL for security vulnerabilities, HIGH for bugs or major code smells, MEDIUM for minor issues, and LOW for style suggestions. CRITICAL and HIGH issues should be fixed before merging.

Can a code review approve changes if a review agent is unavailable?

No. If the code-reviewer or architect agent lane is missing, skipped, or fails, the review reports an unavailable-review state and blocks approval. Self-review is never used as a fallback for the independent lanes.

When does a code review result in REQUEST CHANGES?

The final verdict is REQUEST CHANGES when the architect lane returns BLOCK status, when the code-reviewer lane requests changes, or when required independent review evidence is unavailable. A WATCH status downgrades the result to COMMENT.