Why is dissent preserved in the output report?

Preserving dissent ensures that unresolved disagreements between the AI models are visible, allowing human developers to weigh the different perspectives and make the final decision.

Can I use this for specific git changes?

Yes, by passing the --git-context flag, you can provide specific diffs or recent commit history to the skill for a more targeted and contextual review.

What is the cost of running this review?

A typical review costs between $0.002 and $0.005, which primarily covers the API call for the external cardiologist model.

When is the best time to run a Cardio Review?

It is most effective after implementing significant changes (3+ files), before merging a major feature branch, or when seeking an external perspective on complex architecture.

How does the multi-model consensus system work?

It utilizes a 'Cardiologist' model for primary analysis and a 'Neurologist' model to identify potential blind spots, synthesizing both views into a comprehensive final report.

Multi-Model Consensus Review

Name: Multi-Model Consensus Review
Author: supportersimulator

bysupportersimulator

0•

コラボレーションツール

Performs rigorous cross-examination of code and architectural decisions using a multi-LLM consensus system to identify hidden blind spots.

Cardio Review leverages a unique dual-model architecture to provide deep, external analysis of your codebase and architectural changes. By pairing a primary analyzer (the Cardiologist) with a secondary model dedicated to detecting blind spots (the Neurologist), it ensures that critical decisions are vetted from multiple perspectives. This skill is particularly valuable for complex refactors and significant feature merges, providing a synthesized report that highlights both consensus and healthy dissent to help developers avoid costly oversights and architectural debt.

主な機能

01Automated blind-spot detection to challenge initial findings

02Generation of actionable recommendations for next steps

03Dual-model cross-examination for unbiased code analysis

04Transparent report synthesis that preserves dissenting opinions

050 GitHub stars

06Support for git context to review specific recent changes

ユースケース

01Vetting high-stakes architectural shifts before implementation

02Auditing critical code paths like authentication or data migrations

03Reviewing large feature branches involving multiple file changes

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add supportersimulator/3-surgeons cardio-review

For use in Claude.ai and ChatGPT

主な機能

01Automated blind-spot detection to challenge initial findings

02Generation of actionable recommendations for next steps

03Dual-model cross-examination for unbiased code analysis

04Transparent report synthesis that preserves dissenting opinions

050 GitHub stars

06Support for git context to review specific recent changes

ユースケース

01Vetting high-stakes architectural shifts before implementation

02Auditing critical code paths like authentication or data migrations

03Reviewing large feature branches involving multiple file changes

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add supportersimulator/3-surgeons cardio-review

For use in Claude.ai and ChatGPT