Can I define my own analysis criteria?

Yes, you can specify custom perspectives using the 'perspectives:' or '관점:' prefix followed by your specific focus areas like 'cost' or 'compliance'.

What models does jin-gcc use for analysis?

It prioritizes local Codex and Gemini CLIs for parallel execution. If those are unavailable, it uses multiple specialized Claude agents to provide diverse perspectives.

Do I need to have Gemini or Codex installed?

No. The skill automatically detects their presence; if they aren't installed, it utilizes a robust fallback system using Claude agents to ensure the analysis is still completed.

How does it handle conflicting advice from different AI agents?

The skill includes a synthesis step that specifically identifies 'Conflicts & Trade-offs', presenting the pros and cons of each perspective so you can make an informed decision.

Is the report output structured for documentation?

Yes, it generates a standardized Markdown report including an executive summary, consensus items, trade-off tables, and prioritized action items.

Multi-Perspective Analyzer (GCC)

Name: Multi-Perspective Analyzer (GCC)
Author: cjinzy

bycjinzy

•

开发者工具

Conducts comprehensive, multi-model technical reviews by synthesizing insights from Codex, Gemini, and Claude across diverse specialized domains.

The jin-gcc skill facilitates deep architectural and code reviews by orchestrating a 'council' of AI agents. It attempts to utilize local Codex and Gemini CLI tools for parallel processing, falling back to specialized Claude instances if they are unavailable. The skill analyzes tasks through five core lenses—backend architecture, UX/frontend, security, performance, and side effects—identifying consensus and trade-offs between different technical priorities. This multi-perspective approach ensures that complex engineering decisions are vetted for risks and optimizations before implementation.

主要功能

01Multi-model orchestration using Gemini, Codex, and Claude agents

02Automated synthesis of conflicting viewpoints and trade-offs

03Support for custom analysis perspectives via command prefixes

042 GitHub stars

05Comprehensive reporting with prioritized action items

06Parallel domain-specific analysis (Security, UX, Performance, Architecture)

使用场景

01Automated security and performance auditing of complex PRs

02Conducting pre-deployment architectural reviews and risk assessments

03Evaluating UI/UX proposals against backend technical constraints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cjinzy/jin-claude jin-gcc

For use in Claude.ai and ChatGPT