code-review-perspectives

Reviews code changes across correctness, architecture, security, conventions, and complexity perspectives.

138|5|Updated Sep 19, 2024
One-click install
npx skills add https://github.com/macalbert/envilder --skill code-review-perspectives-macalbert
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-review-perspectives
Source: https://github.com/macalbert/envilder/tree/main/.github/skills/code-review-perspectives
Command: npx skills add https://github.com/macalbert/envilder --skill code-review-perspectives-macalbert

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Code reviews often miss issues because reviewers look at changes from a single angle, letting security flaws, architecture violations, or logic bugs slip through. This Skill structures review into five independent analysis perspectives with a shared severity model, producing a deduplicated, prioritized findings report. ## Core Features & Use Cases - Five Parallel Perspectives: Analyzes correctness, hexagonal architecture boundaries, security (OWASP, secret exposure), project conventions (Biome, test naming, InversifyJS), and CRAP complexity. - Severity Model with Anti-Inflation Rules: Classifies findings as Critical, High, Medium, or Low with strict criteria that prevent severity inflation. - Verification Before Reporting: Confirms findings via test suites, lint checks, and type checks so only verified issues are reported. - Use Case: When reviewing a pull request in a TypeScript hexagonal-architecture project, run all five perspectives on the diff, then receive a prioritized report with file:line evidence and fix directions. ## Quick Start Review the changes in this pull request using the five code review perspectives and report prioritized findings with severity levels.

Frequently Asked Questions about code-review-perspectives

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

FAQPage Schema
How do I review a pull request for security and correctness issues?

Run structured review perspectives over the PR diff: correctness checks for logic bugs and unhandled promises, plus security checks for injection, secret exposure, and unsafe process.env access. Findings are deduplicated, severity-ranked, and reported with file:line evidence.

What code review checklist works for hexagonal architecture projects?

Check that domain code never imports infrastructure, ports and adapters are used correctly, DI wiring is complete, and layer dependencies flow domain to application to infrastructure. Combine this with convention checks for Biome style and InversifyJS decorator usage.

How are severity levels assigned to code review findings?

Findings are classified as Critical, High, Medium, or Low based on objective criteria: Critical requires demonstrable production failure, High requires a concrete input producing incorrect behavior, and convention violations default to Low unless they change runtime behavior.

Can this review approach modify my code automatically?

No, the review is strictly read-only analysis. It never modifies files, only reports confirmed findings with fix directions, and verifies suspicions through test runs and lint checks before including them in the report.

What is CRAP complexity in code review?

CRAP combines cyclomatic complexity with test coverage to flag risky methods. Methods with complexity of 4 or more lacking proportional coverage are flagged, and methods at complexity 6 or above get extraction recommendations.