multi-model-code-review

Coordinate a three-model code-review workflow with cross-model consensus.

Updated Feb 20, 2026

Search Tags:#automation #verification #reconnaissance #multi-model #consensus #code-review

Ready to run

npx skills add https://github.com/pixelsquared/claude-skills --skill multi-model-code-review

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: multi-model-code-review
Source: https://github.com/pixelsquared/claude-skills/tree/main/plugins/code-review/skills/code-review
Command: npx skills add https://github.com/pixelsquared/claude-skills --skill multi-model-code-review

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill orchestrates a multi-model, three-phase code-review workflow to surface verified findings with cross-model consensus, reducing review time and increasing confidence.

Core Features & Use Cases

Three-phase review: Reconnaissance, deep analysis, and verification with model-specific roles to cover structure, behavior, and validation.
Consensus scoring: Automated cross-model comparison to classify findings as Confirmed, Likely, or Possible.
End-to-end reporting: Generates a codebase map, per-dimension findings, and a final consolidated report with optional verification results.

Quick Start

Start a project review by invoking the skill on a repository or path, then follow the prompts to collect phase reports and run the merged analysis.

Frequently Asked Questions about multi-model-code-review

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

FAQPage Schema

How does multi-model code review improve finding accuracy?▼

Multi-model code review improves accuracy by orchestrating a three-phase workflow where independent models cross-check findings during a verification stage, classifying results as Confirmed, Likely, or Possible based on consensus.

How do I run an automated code review on a full project directory?▼

Run automated code review on a full project by invoking the workflow on a repository path, then following prompts to collect reconnaissance, deep analysis, and verification reports to generate a final consolidated map.

What is the difference between reconnaissance and verification in code analysis?▼

In code analysis, reconnaissance maps the structural codebase, while verification validates deep behavioral findings through cross-model consensus, ensuring only verified issues with evidence appear in the final report.

Can I use consensus scoring to review specific file paths instead of a whole repository?▼

Yes, you can apply the consensus scoring workflow to specified paths rather than a full repository, generating phase maps and targeted verification reports for the exact codebase sections you need analyzed.

What outputs should I expect from an automated multi-model code review?▼

Outputs from automated multi-model code review include a codebase map, per-dimension findings from reconnaissance and deep analysis, consensus scores, and a final consolidated report with optional verification evidence.