AI Board Review for Multi-Model Solution Design
Pass your requirements, ideas, or problem statement as arguments
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.
Your raw program ask is transformed into a clear, professional, well-formatted Markdown specification. This ensures consistency and readability for subsequent analysis steps.
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.
The refined requirements are submitted to two distinct AI models for independent review. Each brings different strengths and perspectives to the solution.
High-level reasoning, integration patterns, user flow design
Technical implementation, code feasibility, stack choices
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.
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.
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.
Refined requirements specification
Claude's initial analysis
Gemini's review (UX/integration)
Codex's review (technical)
Synthesized final solution
Executive presentation website
# 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/
Each AI has different training data and reasoning patterns
What one model misses, another may catch
Technical feasibility meets user experience
Full trail of analysis files for reference
The workflow operates under these guidelines:
$HOME/projects/ainotes/createsolution/SYSTEM-OVERVIEW.md but is not limited by it