Which AI models are used for the review process?

The skill prefers using gpt-5.2-codex via the Codex CLI with 'xhigh' reasoning effort, but automatically falls back to Claude Opus if Codex is unavailable.

Can this skill perform external research?

Yes, it enables web search and fetch capabilities to verify technical claims and ensure the plan follows current industry best practices.

What is the primary role of the External Consensus skill?

It acts as an unbiased reviewer that takes three different agent reports and synthesizes them into one final, actionable implementation plan.

How long does the consensus process typically take?

Because it uses maximum reasoning depth for thorough analysis, the skill is configured to allow up to 30 minutes of wall time to complete.

External Consensus

Name: External Consensus
Author: Synthesys-Lab

bySynthesys-Lab

•

협업 도구

Synthesizes balanced implementation plans from multi-agent debate reports using advanced external AI reasoning and web research.

소개

The External Consensus skill serves as a neutral, high-reasoning 'referee' for complex software planning workflows. By aggregating diverse perspectives from multi-agent debates, it utilizes external models like Codex or Claude Opus to resolve architectural conflicts and fact-check technical decisions via web search. This skill is particularly useful when moving from high-level feature brainstorming to concrete implementation, ensuring that the final roadmap is unbiased, follows industry best practices, and addresses all identified risks and trade-offs.

주요 기능

High-reasoning analysis via Codex or Claude Opus
37 GitHub stars
Multi-agent debate synthesis and conflict resolution
Structured risk assessment and mitigation planning
Automated implementation plan and LOC estimation
External web research for technical fact-checking

사용 사례

Automating the transition from multi-perspective critiques to a single actionable plan
Validating architectural decisions against real-world SOTA patterns using web search
Synthesizing a final implementation roadmap from conflicting AI agent proposals

소개

주요 기능

High-reasoning analysis via Codex or Claude Opus
37 GitHub stars
Multi-agent debate synthesis and conflict resolution
Structured risk assessment and mitigation planning
Automated implementation plan and LOC estimation
External web research for technical fact-checking

사용 사례

Automating the transition from multi-perspective critiques to a single actionable plan
Validating architectural decisions against real-world SOTA patterns using web search
Synthesizing a final implementation roadmap from conflicting AI agent proposals