Mcp Christensen

io.github.m2ai-mcp-servers由 GitHub 帳號 m2ai-mcp-servers 發布v0.1.1更新於 Oct 24, 2026

Clayton Christensen persona agent for strategic advisory using disruption theory

已驗證STDIO僅桌面效率與工作流知識與記憶

概覽

AI 產生的概覽

以克雷頓·克里斯汀生人格提供策略顧問的伺服器,讓助理用破壞式創新理論、待辦任務與能力框架分析商業決策。

可以做什麼
  • 對商業決策進行完整的策略分析,可指定框架並補充組織與市場背景
  • 將目前情境與經典案例比對,如鋼鐵迷你鋼廠、磁碟機、奶昔、本田機車與英特爾微處理器
  • 依摘要、完整或僅問題三種詳細程度取得框架說明
  • 以該人格的語氣、診斷式提問與謙遜表述來組織回答,而非確定性邏輯
適用情境
適合希望以克里斯汀生理論為基礎討論策略的場合,例如判斷某項舉措是延續性或破壞性、辨識非消費者,或釐清客戶要完成的任務。它是輔助組織提示的顧問工具,不提供市場資料或計算。
執行需求
以 Node.js 本機程序執行;README 說明需複製儲存庫、執行 npm install 與 npm run build,再把 Claude Desktop 或 Claude Code 指向編譯後的進入點檔案。未宣告帳號、API 金鑰或環境變數。
安裝前請注意
README 說明此伺服器提供的是提示塑造脈絡而非確定性邏輯,輸出屬建議性質,應自行查證。從原始碼安裝需自行建置專案。未描述任何憑證或寫入操作。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Mcp Christensen,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Desktop only,透過 STDIO。 STDIO 服務會啟動本機處理程序,因此需要 SourceWeft 桌面主機。

其他 MCP 客戶端

參照 儲存庫 中的啟動說明。

README

Christensen MCP Server

A Clayton Christensen persona agent for strategic advisory, implemented as an MCP (Model Context Protocol) server.

Apply disruption theory, jobs-to-be-done, and capabilities analysis to your business decisions.

What This Does

This MCP server gives Claude the ability to analyze decisions through Clayton Christensen's frameworks:

  • Jobs-to-Be-Done: What job is the customer hiring your product to do?
  • Disruption Theory: Is this sustaining or disruptive? Who are the non-consumers?
  • Capabilities-Processes-Priorities: Do you have the capability to execute?
  • Resource Dependence: What constraints do your resource providers create?

When you connect this server to Claude Desktop or Claude Code, you get a strategic advisor that thinks like Christensen—asking the right questions, connecting to historical patterns, and offering theory-grounded insights with appropriate humility.

Quick Start

Installation

bash
# Clone the repositorygit clone https://github.com/m2ai-mcp-servers/mcp-christensen.gitcd mcp-christensen
# Install dependenciesnpm install
# Buildnpm run build

Configure Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on Mac or %APPDATA%\Claude\claude_desktop_config.json on Windows):

json
{  "mcpServers": {    "christensen": {      "command": "node",      "args": ["/path/to/christensen-mcp/dist/index.js"]    }  }}

Configure Claude Code

Add to your project's .mcp.json:

json
{  "mcpServers": {    "christensen": {      "command": "node",      "args": ["./path/to/christensen-mcp/dist/index.js"]    }  }}

Tools

analyze_decision

Full strategic analysis using all Christensen frameworks.

Use the analyze_decision tool to evaluate: Should we pursue enterprise customers or focus on SMBs?

Parameters:

  • decision (required): The business decision to analyze
  • context: Additional context about your situation
  • organization: Your org type, size, industry, current focus
  • market: Incumbents, competitors, customer segments
  • focusAreas: Specific frameworks to apply (jtbd, disruption, cpp, resource-dependence)

case_study

Explore Christensen's canonical case studies and match them to your situation.

Use the case_study tool to find patterns matching: We're a low-cost entrant being ignored by incumbents

Available cases:

  • steel_minimills: Low-end disruption with asymmetric motivation
  • disk_drives: New-market disruption with different value network
  • milkshake: Jobs-to-be-done discovery through circumstance
  • honda_motorcycles: Emergent strategy through market learning
  • intel_microprocessors: Capability migration through process investment

get_framework

Learn about any Christensen framework in detail.

Use the get_framework tool with framework: jtbd and detail: full

Frameworks: jtbd, disruption, cpp, resource-dependence, all Detail levels: summary, full, questions-only

Example Conversations

Strategic Decision Analysis

You: I'm trying to decide whether to build individual AI agents or invest in a factory process for creating them. Can you analyze this using Christensen's frameworks?

Claude: Uses analyze_decision tool

Let me think through this using Christensen's frameworks...

Jobs-to-Be-Done: What job are your customers hiring an AI agent to do? If they need specialized expertise (functional), confidence in the advice (emotional), and credibility (social), individual high-quality agents might serve that job better. But if they need to create their own agents quickly, the factory serves a different job entirely.

The Intel Pattern: I'm reminded of Intel's pivot from memory to microprocessors. The insight wasn't about resources—Japanese firms could acquire those. It was about processes. Intel's chip design and manufacturing processes transferred to the new opportunity.

The question I'd ask: What processes have you developed building this first agent that could transfer to a factory? If the process is the real capability, the factory might be more valuable than any single agent.

Case Study Matching

You: We're entering a market with a simpler, cheaper product. Incumbents seem happy to ignore us.

Claude: Uses case_study tool

This matches the Steel Mini-Mills pattern closely.

The integrated steel mills were relieved to cede the rebar market to mini-mills—it was low-margin work. But mini-mills improved and moved upmarket to angle iron, then structural steel, then sheet steel. At each tier, the integrated mills rationally retreated.

The theory would predict: If incumbents are happy to ignore you, that's actually a warning sign for them. The question is whether you're on an improvement trajectory that will eventually satisfy mainstream customers.

What I'd want to understand: Are you improving along dimensions that will matter to their current customers eventually?

How It Works

The MCP server provides prompt-shaping context, not deterministic logic. When you call a tool:

  1. The server structures your decision using Christensen's frameworks
  2. It matches relevant case studies and diagnostic questions
  3. It provides Christensen's voice characteristics and phrases
  4. Claude applies the reasoning with the persona's analytical approach

This design lets Claude do what it does best—nuanced reasoning and judgment—while the MCP tools ensure the analysis stays grounded in Christensen's theory.

Persona Fidelity

The agent is designed to maintain high fidelity to Christensen's actual thinking:

Must include:

  • Reference to specific frameworks (JTBD, disruption, CPP)
  • Questions about the customer's job to be done
  • Humility about predictions ("the theory suggests...")
  • Connection to historical case studies
  • Consideration of organizational constraints

Must avoid:

  • Overconfident predictions
  • Dismissing incumbents as stupid (they're rational!)
  • Generic business advice without theory grounding

Run npm test to validate outputs against these fidelity markers.

Development

bash
# Build TypeScriptnpm run build
# Watch modenpm run dev
# Run validation testsnpm test
# Test with MCP Inspectornpm run inspector

Project Structure

christensen-mcp/├── src/│   ├── index.ts                 # MCP server entry point│   ├── persona-loader.ts        # YAML persona parser│   ├── personas/│   │   └── christensen.yaml     # Core persona definition│   ├── frameworks/│   │   ├── types.ts             # Shared types│   │   ├── jobs-to-be-done.ts   # JTBD framework│   │   ├── disruption.ts        # Disruption theory│   │   └── capabilities.ts      # CPP + Resource Dependence│   ├── tools/│   │   ├── analyze-decision.ts  # Main analysis tool│   │   ├── case-study.ts        # Case study explorer│   │   └── get-framework.ts     # Framework reference│   └── validation/│       ├── fidelity-check.ts    # Output validation│       └── test-decisions.ts    # Test cases├── docs/│   ├── BUILD_LOG.md             # Development process journal│   └── CHRISTENSEN_SPEC.md      # Persona specification└── dist/                        # Compiled JavaScript

Factory Pattern

This agent was built as a proof-of-concept for a persona agent factory. The BUILD_LOG.md documents the entire process so future agents can be built faster.

Key patterns extracted:

  1. Persona YAML structure: identity, voice, frameworks, validation
  2. Framework module pattern: questions, prompts, voice phrases
  3. Fidelity validation: automated testing against persona markers

Credits

Built on Clayton Christensen's work, including:

  • The Innovator's Dilemma (1997)
  • The Innovator's Solution (2003)
  • Competing Against Luck (2016)

This is an educational tool meant to help apply Christensen's publicly available frameworks. It is not affiliated with or endorsed by the Christensen Institute.

License

MIT


Built with Claude Code as a demonstration of MCP server development and persona agent architecture.

來源:README.md,提交 c64b6a3

工具

0
SourceWeft 尚未在沙箱中啟動過該軟體包,因此其工具未知。

版本歷史

1
  1. v0.1.1最新Sep 16, 2026

評分

已滿足適用檢查項的 50%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通25 分

    請保持端點能夠回應 MCP 握手;SourceWeft 每 30 天探測一次。

  • 已宣告認證方式10 分

    請在註冊中心條目中宣告憑證請求標頭,或保持端點能夠回應探測,以便判定其認證方式。

  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list,或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分