ai-berkshire
AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing res
What is ai-berkshire?
AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built for Claude Code / Codex. 4 masters' methodologies + multi-agent adversarial analysis.
Overview of ai-berkshire
What it is
AI Berkshire is a collection of investment research skills designed for Claude Code and Codex. It systematizes the methodologies of four value investing masters: Warren Buffett, Charlie Munger, Duan Yongping, and Li Lu. The framework aims to enable a single person, using AI, to perform professional-grade investment research, described as equivalent to having an entire research team.
Key features
- Four-master methodology: Applies four distinct investment perspectives (Buffett, Munger, Duan, Li) in parallel to produce multi-angle analysis.
- Multi-agent adversarial analysis: The
/investment-teamskill launches four independent AI agents that each search, cross-verify data, and draw conclusions, then a team lead synthesizes the results. - Structured anti-bias mechanisms: Includes an information richness rating (A/B/C), a Mungerian reverse stress test, a rapid disqualification checklist, a contrarian check, and a principle of leaving blanks when data is insufficient.
- Financial accuracy tools: All calculations use Python's
decimal.Decimalfor exact arithmetic, and critical data must be cross-verified from at least two independent sources. - Reproducible research flow: Ensures the same input produces consistent structure and depth of output, allowing horizontal comparison across companies and time.
Use cases
- In-depth equity research: Performing comprehensive value investing analysis on a single listed company.
- Team-based earnings review: Having multiple agents independently analyze a company's earnings report and then synthesize findings.
- Industry screening and funneling: Scanning an entire industry chain for investment opportunities, then filtering from the full market down to a shortlist for deep analysis.
- Portfolio management: Tracking investment theses over time, detecting thesis drift, and managing portfolio concentration and rebalancing.
- Rapid event response: Quickly attributing stock price movements to their underlying causes.
Getting started
Configure the skills for use with Claude Code or Codex. See the repository README for detailed setup instructions and a full list of 20 skills.
License and pricing
This project is open source under the MIT license.
Based on the project's public README and repository metadata.
