Validating Startup Ideas Instantly with cm-brainstorm-idea

As a founder, your greatest asset isn’t your capital, and it isn’t even your code—it’s your time. In the high-velocity world of “Vibe Coding,” where AI can generate hundreds of lines of functional code in seconds, the bottleneck has shifted. The risk is no longer “Can we build this?” but rather “Should we build this, and is this the most efficient way to achieve our business goal?”

We’ve all been there: you have a midnight epiphany for a new feature or a pivot. You open your AI-powered IDE, start prompting, and within an hour, you have a working prototype. But two weeks later, you realize the architecture is a dead end, the user experience is friction-heavy, and the “killer feature” doesn’t actually solve the customer’s core problem. You’ve just fallen into the Execution Trap—building at light speed in the wrong direction.

This is where the cm-brainstorm-idea skill becomes your secret weapon. It is not just a “chatbot for ideas”; it is a disciplined, multi-dimensional strategic gate designed to filter out the noise and refine your vision before a single line of production code is written.

The Core Concept: Moving Beyond “Vibes”

cm-brainstorm-idea is built on the philosophy that Strategic Friction is necessary for long-term Development Velocity. While Vibe Coding emphasizes the flow and the “vibe” of creation, this skill introduces a structured pause. It uses a combination of industry-standard frameworks to stress-test your idea from four critical perspectives:

Technical Feasibility: Can this be built with your current tech stack? What are the hidden complexities (e.g., rate limits, state management, security)?
Product/UX Logic: Does the user flow make sense? Are we following the 48 Laws of UX, or are we creating a “Frankenstein” interface?
Business Strategy: Does this align with your Jobs-To-Be-Done (JTBD)? What is the psychological leverage? Does it solve a “bleeding neck” problem or is it just “nice to have”?
Architectural Integrity: Will this scale, or are we creating a technical debt mountain that will crush us next month?

The TRIZ and Double Diamond Engine

Under the hood, cm-brainstorm-idea employs two powerhouse methodologies: TRIZ (Theory of Inventive Problem Solving) and the Double Diamond.

TRIZ (9 Windows): This framework allows you to view your idea through nine different lenses—looking at the Past, Present, and Future of your Subsystem (the feature), your System (your app), and the Supersystem (the market/ecosystem). It forces you to ask: What did people do before this? How will the ecosystem change once this exists?

Double Diamond: The skill follows a “Diverge then Converge” pattern. First, it expands your idea, looking for all possible variations and edge cases (Divergence). Then, it rigorously prunes those options based on your specific constraints to find the most viable path forward (Convergence).

How It Works: The Strategic Gate

In the Cody Master workflow, cm-brainstorm-idea acts as the Step 0. It sits before cm-planning and cm-execution. If you try to jump straight to coding, you’re guessing. If you start with a brainstorm, you’re engineering.

When you invoke the skill, it doesn’t just say “That’s a great idea!” It triggers a “Strategic Analysis Gate.” It analyzes your current codebase (using the product.md and tech-stack.md context) to ensure the proposal isn’t just a hallucination, but a grounded evolution of your platform.

The output is always structured into 2-3 Qualified Options:

The MVP (Minimal Viable Path): The fastest, lowest-risk way to test the hypothesis.
The Balanced Scale: A robust implementation that balances speed with long-term stability.
The Moonshot: The “Best in Class” version that pushes the boundaries of your current tech.

Practical Example: The “LaunchPad AI” Case Study

Imagine you are the founder of LaunchPad, a platform that helps junior developers find jobs. You want to add an “AI Career Coach” that analyzes a user’s GitHub and gives them a personalized “Skills Gap Report.”

The Old Way (The Trap)

You prompt your AI: “Add an AI coach that reads GitHub and suggests what to learn.” The AI generates a script that fetches GitHub repos, sends them to GPT-4, and dumps a list of keywords into a modal. It works. But it’s slow, expensive, and the users find the feedback generic. You’ve spent three days building a feature that has 10% retention.

The New Way (Using cm-brainstorm-idea)

You invoke the skill: Gemini: use cm-brainstorm-idea to analyze a new "AI Career Coach" feature for LaunchPad.

The skill performs a deep dive and surfaces several critical insights you missed:

Technical Warning: “Fetching full GitHub repo data for every user will hit rate limits instantly and increase latency. Suggest using the GitHub ‘Languages’ and ‘Topics’ metadata first.”
UX Insight: “Junior developers feel overwhelmed by lists of skills. Suggest a ‘Project-Based’ roadmap instead of a keyword list.”
Business Analysis: “The core ‘Job’ here isn’t ‘Learning,’ it’s ‘Confidence.’ The AI should highlight what they already know as much as what they lack.”

The Resulting Qualified Options:

Option A (The 2-Hour MVP): A simple dashboard widget that uses existing profile data to recommend three specific projects to build. No external API calls needed.
Option B (The Robust Implementation): A dedicated “Coach” tab using a vector database to match their GitHub history against real job descriptions.
Option C (The Moonshot): A real-time “Shadow Mentor” that joins their VS Code session and gives advice as they code.

Because you used cm-brainstorm-idea, you decide to start with Option A. You validate it in one afternoon, find that users love the “Project Recommendations,” and then use cm-planning to build the full version. You saved yourself a week of wasted effort on a Moonshot that might not have worked.

Best Practices & Tips for Founders

To get the most out of this skill, you need to approach it like a Board Meeting, not a search query.

1. Provide the “Why,” Not Just the “What”

Don’t just say “I want to add a dark mode.” Say “I want to add dark mode because our analytics show 40% of our users are coding between 11 PM and 3 AM, and we want to reduce eye strain to increase session length.” The more business context you give, the better the TRIZ analysis will be.

2. Embrace the “Devil’s Advocate”

The brainstorm output will often challenge your assumptions. It might tell you an idea is “High Risk / Low Reward.” Don’t ignore this. Use the “Interactive Refinement” phase to ask why it thinks that. The AI has scanned thousands of successful (and failed) architectural patterns; listen to the data.

3. Bridge Immediately to Design

Once you pick an option, don’t just start coding. The brainstorm output is the perfect input for a DESIGN.md file. Use the “Qualified Option” text to feed into cm-ui-preview or stitch to visualize the UI before you touch the logic.

4. The “Pre-Mortem” Iteration

Ask the skill: “Run a pre-mortem on Option B. Why will this fail in 3 months?” This forces the AI to look for scalability bottlenecks, security holes, and “feature bloat” that might not be obvious during the honeymoon phase of a new idea.

Conclusion: Velocity is a Vector

In physics, velocity isn’t just speed; it’s speed in a specific direction. Most founders in the AI era have infinite speed but zero direction. They are running in circles, wondering why they aren’t reaching the finish line.

cm-brainstorm-idea gives you the direction. It transforms you from a “Coder who happens to be a Founder” into a Product Architect. It ensures that every “Vibe” you follow is backed by technical rigor, product-market fit logic, and architectural sanity.

By integrating this skill into your daily workflow, you aren’t just building faster; you’re building smarter. You’re ensuring that when you finally do run cm-execution, the code you’re generating is exactly what your business needs to win.

Stop vibing in the dark. Start brainstorming with intent.