skills.rest
zachmayer
zachmayerCOMMUNITY
Explore all skills by this author →
Website𝕏@AI_Insight_SolsGitHub

discussion-partners

Orchestrate parallel multi-model reviews across Codex, Gemini, and two Claude configurations.

2|Updated Oct 30, 2025
Search Tags:#code review#analysis#orchestration#reasoning#multi-model#consensus
One-click install
npx skills add https://github.com/zachmayer/skills --skill discussion-partners
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: discussion-partners
Source: https://github.com/zachmayer/skills/tree/main/.claude/skills/discussion-partners
Command: npx skills add https://github.com/zachmayer/skills --skill discussion-partners

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill solves the problem of model-specific blind spots and cognitive bias by orchestrating a parallel review process across four distinct AI models to ensure robust, multi-perspective analysis.

Core Features & Use Cases

  • Parallel Multi-Model Review: Simultaneously queries Codex, Gemini, and two Claude configurations (fable and opus) to triangulate answers.
  • Diverse Index Coverage: Leverages unique search indexes from different providers to ensure comprehensive information gathering.
  • Use Case: Use this when facing a complex architectural decision, a difficult debugging session, or a consequential code review where a single model's perspective might be insufficient or biased.

Quick Start

Invoke the discussion-partners skill to perform a four-way review of the current architectural proposal.

Frequently Asked Questions about discussion-partners

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

FAQPage Schema
How do I run a multi-model code review to catch architectural blind spots?▼

A multi-model code review resolves cognitive bias by orchestrating parallel queries across Codex, Gemini, and two Claude configurations to synthesize diverse perspectives on complex problems. This ensures robust triangulation and cross-provider consensus for high-stakes decisions.

Can I use multiple AI models for consensus on difficult debugging sessions?▼

Yes, you can use multiple AI models for consensus on difficult debugging by querying different providers simultaneously. This leverages unique search indexes from different providers to ensure comprehensive information gathering and triangulate findings.

What do I need to configure to perform parallel multi-model analysis?▼

To perform parallel multi-model analysis, you need the Agent tool and configured CLI access for each provider to execute parallel prompts and synthesize findings. This setup enables simultaneous queries across Codex, Gemini, and two Claude configurations.

What's the best way to orchestrate adversarial code reviews across different providers?▼

The best way to orchestrate adversarial code reviews across providers is using a parallel multi-model review process. This applies to high-stakes decision-making and scenarios requiring cross-provider consensus by triangulating findings from distinct AI configurations.

When should I not use a single model for complex architectural decisions?▼

You should not use a single model for complex architectural decisions when facing cognitive bias or insufficient perspectives. A single model's analysis might be insufficient for consequential code reviews, necessitating a multi-model consensus approach instead.

Related Skills

Explore More Software Engineering Skills→
RobinNorbergRobinNorberg
5

cccg

Coordinate parallel analyses from Copilot, Claude, Codex, and Gemini into one synthesis.

Community
Advanced
johnlindquistjohnlindquist
24

think

Apply Gemini-based structured reasoning to complex problems and decisions.

Community
Advanced
bahayonghangbahayonghang
16

codex

Generate, analyze, and research code using Codex CLI commands.

Community
Advanced
laststancelaststance
1

x-agents-cross-review

Coordinate multiple agents to review a scope from diverse perspectives.

Official
Advanced