Is the external review process secure?

Yes, the skill operates in a read-only sandbox mode, meaning the external AI can read your reports and search the web, but it cannot modify your files directly.

Which AI models does this skill utilize?

The skill prefers GPT-5.2 Codex with maximum reasoning effort. If Codex is unavailable, it automatically falls back to Claude Opus.

What is the primary purpose of the External Consensus skill?

It serves as a neutral reviewer that takes three different agent reports (Proposal, Critique, and Reducer) and synthesizes them into one final, actionable implementation plan.

How long does the consensus process take?

Because the skill uses 'xhigh' reasoning depth and performs external web research, it is designed to run for up to 30 minutes to ensure a high-quality synthesis.

What kind of output does the skill provide?

It generates a structured Markdown implementation plan including a consensus summary, codebase analysis, step-by-step instructions, and risk mitigation strategies.

External Consensus Synthesis

Name: External Consensus Synthesis
Author: Synthesys-Lab

bySynthesys-Lab

•

Collaboration Tools

Synthesizes balanced, high-reasoning implementation plans from multi-agent debate reports using external AI reviewers.

The External Consensus skill acts as an unbiased 'tie-breaker' and integrator for complex software planning. After a feature has been debated by multiple AI agents from different perspectives (e.g., proposal, critique, and reduction), this skill invokes a high-reasoning external model like GPT-5.2 Codex or Claude Opus to distill those viewpoints into a single, actionable roadmap. By leveraging advanced reasoning settings and web search capabilities, it ensures that the final implementation plan is not only balanced but also aligned with current industry best practices and security standards.

Key Features

01Utilizes high-reasoning models (Codex/Opus) for deep analysis

02Automated fallback mechanism between Codex and Claude Opus

03Generates structured roadmaps with LOC estimates and risk assessments

0437 GitHub stars

05Incorporates web search for technical fact-checking and SOTA research

06Synthesizes multi-perspective AI debates into a unified plan

Use Cases

01Generating a neutral, high-quality implementation plan for complex features

02Resolving technical conflicts between different AI architectural proposals

03Researching and validating security patterns during the design phase

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add synthesys-lab/agentize external-consensus

For use in Claude.ai and ChatGPT

Download Skill