deep-review

Name: deep-review
Availability: InStock
Author: Coder

Coordinate specialized sub-agents to review code changes and consolidate findings.

2.0k|126|Updated Sep 17, 2025

Search Tags:#parallel #risk-assessment #quality-assurance #code-review #sub-agents #diff-analysis

Ready to run

npx skills add https://github.com/coder/mux --skill deep-review-coder

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: deep-review
Source: https://github.com/coder/mux/tree/main/.mux/skills/deep-review
Command: npx skills add https://github.com/coder/mux --skill deep-review-coder

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill orchestrates high-quality code reviews by splitting the task across focused sub-agents, ensuring multiple perspectives are covered (correctness, tests, consistency, UX, performance, safety).

Core Features & Use Cases

Multi-angle review: Spawns 2–5 specialized sub-agents to evaluate a change from different angles and provide actionable findings with file paths.
Structured synthesis: Aggregates results into a single, coherent report with prioritized issues and recommended fixes.
Use Case: For a PR with complex logic, run parallel reviews to surface edge cases, testing gaps, and UX implications before merge.

Quick Start

Initiate a deep-review on the current change set and request a structured, multi-perspective assessment.

Frequently Asked Questions about deep-review

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

FAQPage Schema

How do I run a parallel code review to surface edge cases before merging a complex PR?▼

A multi-agent code review coordinates 2 to 5 specialized sub-agents to evaluate a change from different angles like correctness, tests, consistency, and UX, aggregating their findings into a single prioritized report with file paths.

What is multi-agent code review and how does it work for diff analysis?▼

A multi-agent code review coordinates 2 to 5 specialized sub-agents to evaluate a diff from different angles like correctness, tests, consistency, and UX, aggregating their findings into a single prioritized report with file paths.

How to get structured findings with file paths from a multi-perspective code review?▼

By orchestrating deterministic sub-agents across UI, backend, and tooling layers, the review generates structured findings with specific file paths, ultimately synthesizing them into a coherent report with recommended fixes.

Can I use parallel sub-agents to assess both backend correctness and UI UX implications in a single pass?▼

Yes, you can spawn parallel sub-agents to assess backend correctness and UI UX implications simultaneously, ensuring multi-perspective validation across your entire change set before merging.

What is the best way to perform risk-assessment and quality-assurance on a complex PR diff?▼

The best way to perform risk-assessment on a PR diff is using a coordinated multi-agent review that splits the task across focused sub-agents to surface testing gaps, edge cases, and safety issues.

How many specialized sub-agents are spawned during a parallel code review?▼

A parallel code review typically spawns 2 to 5 specialized sub-agents to evaluate a change from different angles, ensuring multi-perspective validation and structured synthesis of findings.