xbtlin/ai-berkshire

What is it?

Key features

• 01Provides 22 slash-command skills covering deep research, earnings review, industry screening and portfolio tracking.
• 02Runs a multi-agent team mode where four perspective agents research one company in parallel.
• 03Forces explicit verdicts and price ranges instead of balanced summaries.
• 04Includes a Python tool for market-cap checks, valuation ratios and multi-source cross-validation using exact decimal arithmetic.
• 05Installs for both Claude Code and Codex with provided scripts.

Use cases

• Researching a listed company's business quality, moat and valuation before deciding whether to dig deeper.
• Reviewing an annual or quarterly report line by line with a structured earnings workflow.
• Screening a sector or index for companies that pass a value-investing checklist.
• Tracking whether the thesis behind an existing holding still holds over time.

Quick start

1. 1

   Install Claude Code via npm

   npm install -g @anthropic-ai/claude-code
2. 2

   Clone the repository

   git clone https://github.com/xbtlin/ai-berkshire.git
3. 3

   Install the skills into the Claude Code commands directory from the repo folder

   ./scripts/install-claude-commands.sh
4. 4

   Run a research skill inside Claude Code

   /investment-research 腾讯

Things to know

• !Deep research skills consume many tokens because they run multiple research rounds and multi-agent synthesis.
• !Claude Code or Codex must be installed first, and the skills depend on those clients.
• !The README's track-record figures are self-reported and not independently verified.

In collections

Topics

FAQ

What is ai-berkshire?

A set of Claude Code and Codex skills that run value-investing research on public companies, applying Buffett, Munger, Duan Yongping and Li Lu's frameworks with multi-agent analysis. AI Berkshire is a collection of investment research skills that work with Claude Code and Codex. Each skill is a command such as /investment-research or /earnings-review that walks an AI agent through a fixed workflow for analyzing a listed company, covering business model, moat, management, risks and valuation. The methodology is organized around four value investors: Warren Buffett, Charlie Munger, Duan Yongping and Li Lu, who are assigned different angles so their conclusions can conflict and be checked against each other.

Is ai-berkshire free to use? What license does it have?

ai-berkshire is open source under the MIT license. Check the license terms before using it in commercial products.

Is ai-berkshire still maintained?

Actively maintained. The last commit to ai-berkshire was on Nov 4, 2026.

What can I use ai-berkshire for?

Researching a listed company's business quality, moat and valuation before deciding whether to dig deeper. Reviewing an annual or quarterly report line by line with a structured earnings workflow. Screening a sector or index for companies that pass a value-investing checklist. Tracking whether the thesis behind an existing holding still holds over time.

How do I install ai-berkshire?

Install Claude Code via npm: npm install -g @anthropic-ai/claude-code Clone the repository: git clone https://github.com/xbtlin/ai-berkshire.git Install the skills into the Claude Code commands directory from the repo folder: ./scripts/install-claude-commands.sh Run a research skill inside Claude Code: /investment-research 腾讯

Where can I get ai-berkshire?

The source code and installation instructions are on GitHub: https://github.com/xbtlin/ai-berkshire