deep-review

Name: deep-review
Availability: InStock
Author: latticeHQ

Coordinate multi-angle code reviews using 2–5 specialized sub-agents.

Updated Aug 23, 2026

Search Tags:#analysis #quality #parallel #review #code #sub-agents

Ready to run

npx skills add https://github.com/latticeHQ/latticeWorkbench --skill deep-review-latticehq

💬 Prompt for ChatGPT / Claude Chat▼

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

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinate multi-angle code reviews using specialized sub-agents.

Core Features & Use Cases

Spawn 2–5 sub-agents with focused mandates (e.g., Correctness & edge cases, Tests & verification, Consistency & architecture, UX & accessibility, Performance & reliability, Docs & developer experience).
Return actionable findings with file paths to issues for quick remediation.
Enable rapid, governance-aware reviews by distributing workload and maintaining clear artifact traces.

Quick Start

Spawn the right set of 2–5 sub-agents tailored to the change scope and capture their findings into a consolidated review.

Frequently Asked Questions about deep-review

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

FAQPage Schema

How do I perform a multi-angle code review for correctness, performance, and safety?▼

Multi-angle code review is performed by spawning 2–5 specialized sub-agents with focused mandates covering correctness, tests, consistency, UX, performance, and safety. Each sub-agent analyzes the code changes in parallel and returns actionable findings with specific file paths for quick remediation.

What is the best way to scale code reviews for large changes across different architectural concerns?▼

The best way to scale code reviews for large changes is distributing the workload across multiple sub-agents. This approach maintains governance-aware artifact traces by assigning focused mandates like architecture consistency and reliability to individual agents, capturing their findings into a consolidated review.

Can I use parallel sub-agents to review code changes for UX and accessibility?▼

Yes, you can use parallel sub-agents to review code changes for UX and accessibility. You spawn a sub-agent with a dedicated UX and accessibility mandate alongside other agents for correctness or performance, returning actionable findings with file paths for quick remediation.

How does sub-agent coordination work for automated code analysis?▼

Sub-agent coordination for automated code analysis works by spawning 2–5 agents with focused mandates tailored to the change scope. Each agent independently analyzes its designated area, such as edge cases or developer experience, and returns findings that are consolidated into a single review.

Do I need specific dependencies to run parallel code reviews with sub-agents?▼

No specific dependencies are required to run parallel code reviews with sub-agents. The process relies on spawning 2–5 sub-agents with focused mandates tailored to your change scope, capturing their findings into a consolidated review without external component requirements.

When should I not use a multi-agent approach for code analysis?▼

You should not use a multi-agent approach for code analysis when the change scope is too small to benefit from parallel mandates. The approach is designed to scale to changes of any size by distributing workload, but minimal changes may not require spawning 2–5 sub-agents.