How Duan Yongping's Perspective on the Right Business Powers AI Berkshire's Investment Analysis

AI Berkshire hard-codes Duan Yongping's "right business" philosophy into reproducible agent workflows, evaluating every investment against five core criteria—differentiation, moat, pricing power, light assets, and user-centricity—through dedicated skills and automated research pipelines.

AI Berkshire (xbtlin/ai-berkshire) is an open-source investment research framework that operationalizes the wisdom of four investment masters into executable code. Duan Yongping's perspective on the right business serves as the foundational analytical lens, ensuring that every company evaluation begins with a rigorous assessment of business model quality before financial metrics are considered. This methodology is permanently embedded in the repository's skill layer, making the analysis consistent and auditable across every research report.

The Five Pillars of Duan Yongping's Right Business Framework

According to the source code in skills/dyp-ask.md, Duan Yongping defines the right business through five immutable characteristics. These criteria form the "right-business" checklist that AI Berkshire applies to every investment candidate.

Differentiation and Economic Moat (护城河)

The framework requires identifying differentiation (差异化) and moat (护城河) as the primary defensive characteristics. The agent analyzes whether the business possesses unique competitive advantages that protect market position and profitability from erosion.

Pricing Power and Capital Structure

Pricing power (定价权) and light asset (轻资产) structure determine financial resilience. The analysis verifies whether the company can raise prices without losing customers while maintaining minimal capital expenditure requirements to sustain operations.

User-Centric Orientation

The final pillar, user-centricity (用户导向), demands that profit emerge as a natural consequence of solving user problems rather than through extraction tactics. This philosophical anchor prevents the model from favoring businesses with unsustainable revenue models.

How AI Berkshire Implements Duan Yongping's Right Business Perspective

The repository implements Duan Yongping's perspective through a three-layer architecture that ensures the right business analysis executes consistently across all research workflows.

The Standalone Skill Layer (skills/dyp-ask.md)

The /dyp-ask command exposes Duan's framework as a standalone interactive skill. Users can query specific business models directly, and the agent responds using the exact five-pillar structure defined in the markdown source. This file contains the canonical definitions and tone guidelines that govern all Duan-style responses.

The Integrated Research Pipeline (skills/investment-research.md)

Step 2 of the investment research workflow ("生意本质分析 — 段永平'对的生意'") mandates the right business analysis as the first analytical module. When invoking /investment-research <ticker>, the agent generates a one-sentence business definition followed by the five-pillar checklist. The workflow includes a Duan-style追问 (段永平式追问) that forces the model to articulate the core insight in a single sentence: "这门生意好在哪？如果只能用一句话描述..."

Data Validation and Agent Orchestration

The tools/financial_rigor.py module supplies the precise data feeding Duan's analysis, verifying revenue composition, margin trends, and capital intensity metrics that validate or challenge the qualitative assessment of the five pillars. The AGENTS.md documentation defines how the Skill layer interacts with the underlying Codex and Claude infrastructure, clarifying how the Duan skill is orchestrated by the Team Lead agent to ensure proper execution context.

Practical Usage and Code Examples

AI Berkshire exposes Duan Yongping's right business framework through both interactive commands and programmatic APIs.

Quick Business Model Assessment

Invoke the dedicated Duan skill to analyze any business model immediately:

# Direct query to Duan's perspective

/dyp-ask 这家公司的业务模式是什么？

The agent returns a structured analysis following the five criteria:

好生意的特征：
- 差异化：平台型双边网络效应
- 护城河：用户锁定 + 转换成本高
- 定价权：能在不失客的情况下提价
- 轻资产：主要是软件/平台，无大量资本支出
- 用户导向：以用户需求为中心，利润自然产生

This output derives directly from the rules encoded in skills/dyp-ask.md.

Full Investment Research Integration

Run a comprehensive research report that includes Duan's right business analysis as the foundational module:

# Complete research workflow for Tencent

/investment-research 腾讯

The workflow automatically executes:

1. Business Model Module (Duan): Generates the one-sentence definition and five-pillar checklist
2. Moat Analysis (Buffett): Evaluates competitive durability
3. Reversal Check (Munger): Assesses psychological biases
4. Civilization Trend (Li Lu): Positions within macro trends

The Duan section appears under "生意本质分析" in the final markdown report.

Programmatic Access via Python

Access the right business analysis through the Codex API:

import codex

# Execute investment research with Duan's module active

report = codex.run_skill(
    "investment-research",
    args=["AAPL"]  # Apple Inc.

)

print(report)  # Contains right-business analysis as first module

The investment-research skill pulls Duan's logic directly from skills/dyp-ask.md, ensuring identical analytical rules apply whether called via CLI or Python API.

Summary

Duan Yongping's perspective on the right business is permanently embedded into AI Berkshire's DNA through:

• Canonical definitions in skills/dyp-ask.md that codify the five-pillar framework (differentiation, moat, pricing power, light assets, user-centricity)
• Mandatory execution in skills/investment-research.md Step 2, ensuring every research report begins with business model quality assessment
• Reproducible workflows that trigger the same logical rules across CLI and API interfaces
• Data validation through tools/financial_rigor.py, which supplies financial metrics to verify qualitative pillar assessments

This architecture guarantees that Duan's philosophy—the right business is the foundation of every investment decision—is applied consistently and auditably to every evaluation.

Frequently Asked Questions

What are the five specific criteria of Duan Yongping's right business framework in AI Berkshire?

The five criteria are differentiation (差异化), economic moat (护城河), pricing power (定价权), light asset structure (轻资产), and user-centric orientation (用户导向). These are hard-coded in skills/dyp-ask.md and serve as the mandatory checklist for every business model analysis in the repository.

How does the right business analysis integrate with other investment masters in AI Berkshire?

The Team Lead agent spawns four master agents (Buffett, Munger, Duan, and Li Lu) when executing /investment-research. Duan's agent runs first, establishing the business model foundation through the right business framework. This output feeds into subsequent modules—Buffett's moat analysis, Munger's reversal checks, and Li Lu's civilization trends—creating a layered evaluation where business quality precedes financial metrics.

Can developers customize the right business criteria for specific industries?

Yes. Since the framework lives in standard markdown files (skills/dyp-ask.md and skills/investment-research.md), developers can modify the five-pillar definitions or add industry-specific nuances without altering the underlying Python code. The agent layer reads these files dynamically, making the system adaptable while maintaining the core Duan Yongping philosophy.

What financial data validates the right business qualitative assessment?

The tools/financial_rigor.py module supplies quantitative verification for Duan's qualitative pillars, including revenue mix analysis, margin trend validation, and capital expenditure ratios. This ensures that claims about pricing power, light assets, or differentiation are grounded in actual financial performance rather than narrative alone.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}