code-review

Automate code quality reviews for security, reliability, performance, and taste.

1.8k|62|Updated Jan 11, 2026
One-click install
npx skills add https://github.com/LibPDF-js/core --skill code-review-libpdf-js
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/LibPDF-js/core/tree/main/.opencode/skills/code-review
Command: npx skills add https://github.com/LibPDF-js/core --skill code-review-libpdf-js

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill automates brutally honest code quality reviews focusing on security, reliability, performance, and taste.

Core Features & Use Cases

  • Brutal, honest evaluation of code quality across security, reliability, performance, and taste
  • Structured findings across multiple passes (taste, safety, reliability, clarity)
  • Aggregates results into a single report for developers and product teams
  • Applicable to code reviews, pull requests, and architectural reviews

Quick Start

Use this skill against a repository or PR by specifying the scope and target files. The skill will run parallel passes and generate a consolidated report at .agents/scratch/code-review-<scope>.md

Frequently Asked Questions about code-review

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

FAQPage Schema
How do I automate brutally honest code quality reviews for a pull request?

To automate brutally honest code quality reviews, run this skill against a repository or PR to execute parallel passes across taste, safety, reliability, and clarity, aggregating the results into a consolidated markdown report.

What is multi-pass code review and how does it find security and reliability issues?

Multi-pass code review uses parallel subagents to independently analyze different aspects of a codebase, such as security and reliability, ensuring structured issue discovery across the project before consolidating findings into a report.

Can I use automated code review for architectural and performance evaluations?

Yes, you can use automated code review for architectural and performance evaluations by specifying the target scope and files. The skill analyzes the codebase to surface concrete findings related to performance and architecture.

Does this automated code review tool require any external dependencies?

No, this automated code review tool requires no external dependencies. It operates independently to evaluate code quality, security, and performance, directly generating a consolidated report in your local environment.

Where does the automated code review output its consolidated findings report?

The automated code review outputs its consolidated findings report to a markdown file located at `.agents/scratch/code-review-<scope>.md`, containing aggregated results from the multi-pass analysis.