skills.rest
austintheriot
austintheriotCOMMUNITY
Explore all skills by this author →
GitHub

expert-review

Classify code regions and dispatch specialist evaluators to produce ranked review reports.

1|Updated Jul 31, 2023
Search Tags:#code review#bug detection#read-only analysis#multi-expert panel#severity ranking#diff mode#security lens
One-click install
npx skills add https://github.com/austintheriot/dotfiles --skill expert-review-austintheriot
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: expert-review
Source: https://github.com/austintheriot/dotfiles/tree/main/.claude/skills/expert-review
Command: npx skills add https://github.com/austintheriot/dotfiles --skill expert-review-austintheriot

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

expert-review eliminates missed defects and design-quality gaps by turning a single code change into a structured, multi-lens panel assessment with severity-and-confidence ranking.

Core Features & Use Cases

  • Multi-expert regional classification: Breaks the code into regions and routes each region to the most relevant specialist lenses (bugs, security, performance, concurrency, readability, API design, etc.).
  • Diff or full survey modes: Reviews either changed regions (diff) or a named path/feature in full (survey), focusing attention according to scope.
  • Synthesis that ranks outcomes: Produces one combined report with severity ordering and confidence that reflects cross-expert agreement.
  • Safety stance: read-only: Intended for genuine review passes without applying fixes or posting to external systems.

Quick Start

Run expert-review in diff mode for your current changes to obtain a ranked panel report covering bugs, security, correctness, and maintainability.

Frequently Asked Questions about expert-review

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

FAQPage Schema
How do I get a severity-ranked code review for a pull request diff?▼

To get a severity-ranked code review, run expert-review in diff mode on your current changes to obtain a combined panel report covering bugs, security, correctness, and maintainability with cross-expert confidence ranking.

What is a multi-expert code review and how does it classify bugs?▼

A multi-expert code review breaks code into regions and routes each to specialist lenses for bugs, security, performance, and concurrency, synthesizing the panel output into a single severity-ranked report.

Can I use this to do a full code survey instead of just reviewing a diff?▼

Yes, you can run a full code survey instead of a diff review by targeting a named path or feature, allowing the expert panel to focus its attention across the entire specified scope.

Does the code review process modify my code or apply fixes automatically?▼

No, the code review process is strictly read-only, meaning it performs genuine analysis passes to identify defects and design gaps without applying fixes or posting to external systems.

What lenses are required for a comprehensive bug detection pass?▼

A comprehensive bug detection pass requires a bug-hunting lens, at least one functional-programming lens, and a first-principles scope check to ensure correctness across the provided code regions.

What is the best way to review code correctness and security across different languages?▼

The best way to review code correctness across languages is using a multi-expert panel that dispatches specialist evaluators for security and concurrency, ranking outcomes by severity and cross-expert agreement.

Related Skills

Explore More Software Engineering Skills→
msd-dev-labmsd-dev-lab

codex-review

Automate iterative code reviews with read-only Codex CLI and Claude Code.

Official
Advanced
an8079an8079
4

auto-critique

Automate code review for quality, correctness, security, and performance.

Community
Advanced
salemazielsalemaziel

testing

Generate unit and integration tests from code analysis across multiple frameworks.

Community
Advanced
austintheriotaustintheriot
1

agent-modes

Standardize output formats for review, plan, and consult agent workflows.

Community
Basic