Introducing: Curiosity Stack

A structured research thinking tool for Claude Cowork. Decode any technology or market signal into 2nd and 3rd order business insights.

Six layers, one method, infinite topics

For years I’ve been asked some version of the same question: how do you find a business with real potential before it is on everyone’s radar? Before it is widely discussed on X, before the FinFluencers and analysts are writing about it, before retail investors are talking about it?

The honest answer is that it is not instinct. It is a method. And I have been refining it for a long time.

I call it the Curiosity Stack.

Today I’m releasing it as a Claude Cowork plugin — free to install, free to use.

What it is

The Curiosity Stack is a six-layer question framework for going deeper than headlines.

Most people hear about a technology and stop at Layer 1. They read that AI hallucinates, form a vague impression, and move on. The Curiosity Stack keeps going:

L0 — Signal. What is everyone talking about, and why now?

L1 — Mechanics. What is this really? Not the PR version but the actual technical or economic reality.

L2 — Cause Tree. What causes this to happen? Not one cause but three to five. Which are structural? Which are cyclical?

L3 — Solution Space. For each cause, what are the solutions? Each solution category is an industry.

L4 — Build Requirements. What does each solution actually require? Inputs, infrastructure, talent, data. What is scarce?

L5 — Value Chain Actors. Who does this work? Named companies global leaders and Indian proxies.

L6 — Research Landscape. How is this space typically accessed? Listed, unlisted, pre-IPO? What is the timing?

Here is how the hallucination example plays out:

AI hallucinates → root cause: bad training data → solution: human annotation and data labeling → what is required: annotation tools, quality control, domain expertise → who does this: a specific niche of companies at the data layer that most retail investors have never considered

You started with a headline. You ended with a specific layer in the value chain with businesses that have genuine potential and almost no mainstream coverage.

That is Layer 1 vs Layer 5 thinking. The framework makes the path explicit and repeatable.

What the plugin actually does

The plugin walks you through the decomposition in a Cowork session, one layer at a time, like a hand holding teacher. It asks before it tells. You build the understanding by answering, not by reading a wall of text.

After each session you pick one output format:

• Value Chain diagram: a structured map of the full decomposition

• Interactive Mindmap: visual, navigable, shareable

• Research Note: written summary you can save or share

Note: Value Chain & Interactive Mind Map tasks can take a bit longer than you prefer. Please be patient as the plugin works through information & generates the collaterals. Once generated, you are free to use them for your own benefit.

Here is a Mind Map Example:

Every session also automatically generates:

• A Thesis Stress Test with factors proponents of this thesis cite, factors critics cite, the core assumption the whole thesis rests on, and how badly wrong it could be

• 5 curated reading recommendations targeted at the specific topic and the layers where you have gaps

• An option to save everything back to your Drive, Notion, or Airtable

The plugin connects to 17 sources: Google Drive, Gmail, Notion, Slack, Airtable, OneDrive, Confluence, GitHub, and more. Your existing research surfaces automatically when relevant. You’re not starting from zero every session.

Who can use it

Two types of people:

Enterprise professionals who want to understand where AI and technology are actually going & not just the headline, but the business implications, the value chain, the Indian angle. The kind of understanding that makes you the most useful person in any room discussing these topics.

Retail investors looking for non-obvious businesses with real potential before they are covered, before they are priced in, before the wave hits mainstream awareness.

One method. Two use cases. The same discipline applies.

Important to note

This is a research and structured thinking tool.

I am not a SEBI registered investment advisor. Nothing this plugin produces is investment advice, a buy/sell recommendation, or a signal to act on. It helps you think and structure your research and what you do with that thinking is entirely yours.

The plugin is built with this baked in. It will not give you price targets. It will not tell you what to buy. It will help you understand deeply, structurally, with specific names at the end of the chain.

How to get it

Claude Cowork → Browse Plugins → search “Curiosity Stack”

Or download directly from GitHub and upload to Cowork.

Note: The plugin is currently under review with Anthropic. If you prefer to wait and try only after it is approved, for security reasons, please be patient. I will inform you if it is approved. If not, I will fix the findings are submit again. Until then, the choice is yours - whether to download it & upload locally or just wait for anthropic to approve.

Run /curiosity-stack:setup first — it takes about two minutes to connect your notes and set your context. Then run /curiosity-stack:decompose on any topic you’ve been meaning to understand properly.

A final note

There’s a line I keep coming back to:

“The intent is not to reach the end of the process but to go through the process itself.”

The structured understanding forming in your head as you walk the layers that is the actual asset. Not the diagram. Not the research note. Not the company names. The mental model that now shapes every future observation you make.

That’s what this is for. Follow this & you will keep discovering new business ideas in the future.

Try it. Tell me what topic you ran first.

And, if you don’t find value, it’s fair. Please be kind when you submit the feedback. I am also a learner & can make mistakes.

— Ameya