← Back to User-Level Skills

Create Solution Workflow

AI Board Review for Multi-Model Solution Design

How to Invoke

/createsolution {your program ask here}

Pass your requirements, ideas, or problem statement as arguments

Overview

The createsolution skill automates a comprehensive solution design workflow. It takes your program ask (requirements, ideas, or problem statement), refines it into a professional specification, submits it to multiple AI models for diverse perspectives, synthesizes the feedback into a consolidated solution, and generates an executive presentation website.

This workflow ensures you get well-considered solutions that benefit from multiple AI perspectives while maintaining a structured, auditable design process.

Workflow Diagram

1
Refinement
Format requirements
2
Analysis
Initial proposal
3
AI Review
Gemini + Codex
4
Synthesis
Merge best ideas
5
Optimize
Critical review
6
Present
Web summary

Step-by-Step Process

1 Refinement & Formatting

Your raw program ask is transformed into a clear, professional, well-formatted Markdown specification. This ensures consistency and readability for subsequent analysis steps.

$HOME/projects/ainotes/createsolution/{descriptor}.md

2 Initial Analysis (Claude)

Claude (the orchestrating AI) analyzes the refined requirements and proposes a primary solution. This serves as the baseline proposal that will be reviewed by other models.

$HOME/projects/ainotes/createsolution/{descriptor}.claude.md

3 AI Board Review

The refined requirements are submitted to two distinct AI models for independent review. Each brings different strengths and perspectives to the solution.

Gemini

High-level reasoning, integration patterns, user flow design

{descriptor}.gemini.md
💻

Codex (ChatGPT)

Technical implementation, code feasibility, stack choices

{descriptor}.codex.md

4 Synthesis (First Pass)

All three responses (Claude, Gemini, Codex) are analyzed together. The best components from each are identified, conflicts are resolved, and everything is merged into a consolidated solution document.

$HOME/projects/ainotes/createsolution/{descriptor}.final.md

5 Final Optimization

A second critical pass reviews the final.md file for gaps, inefficiencies, or overlooked considerations. Improvements are incorporated and the file is updated in place.

$HOME/projects/ainotes/createsolution/{descriptor}.final.md (updated)

6 Executive Presentation

A web-ready summary is created and deployed to the Nginx server. This includes an executive summary, high-level flow diagrams, and a walkthrough of the solution for easy stakeholder comprehension.

$HOME/projects/nginx/sites/{descriptor}/index.html

Output Files Generated

📝

{descriptor}.md

Refined requirements specification

🤖

{descriptor}.claude.md

Claude's initial analysis

{descriptor}.gemini.md

Gemini's review (UX/integration)

💻

{descriptor}.codex.md

Codex's review (technical)

{descriptor}.final.md

Synthesized final solution

🌐

sites/{descriptor}/

Executive presentation website

Example Usage

Invoking the Skill

# Simple invocation with inline ask
/createsolution I want to build a system that tracks my stock options,
monitors price movements, sends alerts when thresholds are hit,
and provides a dashboard for portfolio visualization.

# The skill will automatically:
# 1. Create optionstracker.md (refined requirements)
# 2. Create optionstracker.claude.md (initial analysis)
# 3. Create optionstracker.gemini.md (Gemini review)
# 4. Create optionstracker.codex.md (Codex review)
# 5. Create optionstracker.final.md (synthesized solution)
# 6. Deploy to nginx.ai-servicers.com/optionstracker/

Why Multi-Model Review?

🔍

Diverse Perspectives

Each AI has different training data and reasoning patterns

🛡

Catch Blind Spots

What one model misses, another may catch

Balanced Solutions

Technical feasibility meets user experience

📊

Auditable Process

Full trail of analysis files for reference

Technology Constraints

The workflow operates under these guidelines: