MBB Management Consultant AI Skill and Plugin for LLMs

The most comprehensive open-source consulting intelligence layer for any LLM. A structured, hypothesis-driven consulting knowledge base — works with Claude, Codex, Gemini, GitHub Copilot, Cursor, Ollama, and any other AI assistant. Built for MBA students, case interview candidates, business professionals, and anyone who needs to think at McKinsey / Bain / BCG level.

TL;DR: Install this skill and your AI assistant — Claude, Codex, Gemini, Copilot, Ollama, or any other LLM — works like an MBB-grade management consultant, coaching you through case interviews, applying strategy frameworks, sizing markets, and turning messy business problems into board-ready answers. 129 reference modules distilled from public consulting knowledge. Free and open-source; contributions welcome.

Install

Claude Code — one command:

/plugin marketplace add DogInfantry/claude-skill-management-consultant-B1
/plugin install management-consultant

Works in any other AI agent too — Codex, GitHub Copilot, Gemini, Cursor, and Ollama auto-load the skill from AGENTS.md / GEMINI.md / .github/copilot-instructions.md / Modelfile, and any LLM can use it by attaching skill/SKILL.md. Full walkthrough and manual setup: Quick Start · INSTALL.md.

What Is This?

The MBB Management Consultant AI Skill is a structured knowledge system that installs into your AI assistant — Claude, Codex, Gemini, GitHub Copilot, Ollama, and more — giving it the cognitive architecture of a senior strategy consultant. It is not a simple prompt — it is 129 reference modules, organized around 5 knowledge pillars, covering every dimension of consulting work from MECE issue trees and Pyramid Principle communication to M&A due diligence, GenAI enterprise strategy, and case interview coaching.

Unlike generic AI prompts, this skill:

• Applies hypothesis-driven thinking — forms a point of view before gathering data
• Structures every problem MECE (Mutually Exclusive, Collectively Exhaustive)
• Leads every output with the governing thought (Pyramid Principle)
• Draws from 10+ industry reference files and 129 deep-knowledge modules
• Follows the MBB operating model — Frame → Structure → Analyse → Synthesise → Deliver

Compatible with: Claude (Code, Free, Pro, Teams), OpenAI Codex, Google Gemini, GitHub Copilot, Cursor, Ollama (local models), and any other LLM you can paste context into.

How It Works

flowchart LR
    A["🔍 Problem Brief"] --> B["1. Frame"]
    B --> C["2. Structure"]
    C --> D["3. Analyse"]
    D --> E["4. Synthesise"]
    E --> F["5. Deliver"]

    B -.->|reads| G[("129 Reference Files")]
    C -.-> G
    D -.-> G
    E -.-> G
    F -.->|"Board-ready output"| H["✅ Client Value"]

Frame → Structure → Analyse → Synthesise → Deliver. Every problem flows through the same consulting operating system, drawing from 5 knowledge pillars and 129 reference modules.

Knowledge Architecture

The 129 reference files are organized across five knowledge pillars:

mindmap
  root((129 Reference Modules))
    Problem-Solving
      MECE Issue Trees
      Hypothesis Trees
      First Principles
      Mental Models
      McKinsey 7-Step Method
      Decision Under Uncertainty
    Communication
      Pyramid Principle
      Storylining
      Slide Craft
      Executive Presence
      Structured Verbal
      Board Decks
    Case Coaching
      All MBB Case Formats
      Market Sizing
      Case Pattern Library
      PEI STAR Coaching
      Scoring Rubrics
    Industry Intelligence
      Healthcare & Life Sciences
      Financial Services
      Energy & Utilities
      Technology & TMT
      Consumer & CPG
      Industrial & Mfg
      Public Sector
      Real Estate
      Hospitality & Travel
      Emerging Markets
    Specialist Methods
      M&A and Due Diligence
      Post-Merger Integration
      Corporate Restructuring
      GenAI Enterprise Strategy
      Pricing Architecture
      Org Design
      Climate Strategy
      Transformation PMO
      Sales Force Effectiveness
      Process Mining

What This Is

This repository contains a consulting knowledge base for AI assistants — Claude, OpenAI Codex, Google Gemini, GitHub Copilot, Ollama, and any other LLM. When installed, it gives the assistant a structured senior-consultant workflow, built around MBB-style problem-solving and communication patterns built by distilling publicly available MBB frameworks, case prep resources, and consulting methodology into a single knowledge base.

This is not a simple prompt. It is a structured knowledge system with 129 reference files covering every dimension of consulting work, from MECE issue trees and Pyramid Principle communication to due diligence, post-merger integration, GenAI enterprise strategy, and case interview coaching.

It is completely free.

The Core Problem-Solving System

Every consulting engagement — and every case interview — runs on the same two tools. This skill makes your AI assistant fluent in both.

MECE Issue Trees: Diagnosing Any Business Problem

The MECE issue tree is the workhorse of consulting analysis. It decomposes a business problem into branches that are Mutually Exclusive (no overlaps) and Collectively Exhaustive (no gaps), so every potential root cause is surfaced and tested.

flowchart TD
    P["🔴 Why is EBITDA Margin Declining?"]:::root --> R["📉 Revenue Side"]
    P --> C["📈 Cost Side"]

    R --> RV["Volume declining?"]
    R --> RP["Price eroding?"]
    R --> RM["Mix shifting?"]

    C --> CG["COGS increasing?"]
    C --> CS["SG&A bloating?"]
    C --> CO["Other costs?"]

    RV --> RV1["Fewer new customers"]
    RV --> RV2["Higher churn rate"]
    RP --> RP1["Discounting depth"]
    RP --> RP2["Competitor pricing pressure"]

    CG --> CG1["Raw material inflation"]
    CG --> CG2["Labour cost increases"]
    CS --> CS1["Headcount growth"]
    CS --> CS2["Marketing spend"]

    classDef root fill:#dc2626,color:#fff,stroke:#991b1b

Ask Claude to build an issue tree for any business problem. It will always structure branches MECE and prioritize the highest-impact paths for analysis.

The Pyramid Principle: Structuring Every Output

Every deliverable — memo, deck, verbal update, email — follows the Pyramid Principle (Barbara Minto, McKinsey). Lead with the answer. Support with structured reasoning. Evidence comes last.

flowchart TD
    GT["💡 GOVERNING THOUGHT\nState the answer first"]:::top --> W1["Reason 1"]:::mid
    GT --> W2["Reason 2"]:::mid
    GT --> W3["Reason 3"]:::mid

    W1 --> D1["Evidence"]:::base
    W1 --> D2["Evidence"]:::base
    W2 --> D3["Evidence"]:::base
    W2 --> D4["Evidence"]:::base
    W3 --> D5["Evidence"]:::base
    W3 --> D6["Evidence"]:::base

    classDef top fill:#1d4ed8,color:#fff,stroke:#1e3a8a
    classDef mid fill:#2563eb,color:#fff,stroke:#1d4ed8,opacity:0.85
    classDef base fill:#93c5fd,color:#1e3a8a,stroke:#2563eb

The wrong pattern: describe context → walk through analysis → arrive at the answer. The consulting pattern: state the answer → group your reasons → back each with evidence.

Knowledge Sources & IP Notice

This project contains no proprietary, confidential, or insider material from any consulting firm or client engagement. Everything here is derived from publicly available sources only.

What this knowledge base does NOT contain:

• ❌ No client names, engagement details, or project-specific data
• ❌ No internal firm templates, proprietary scoring rubrics, or methodology documents obtained under NDA
• ❌ No unpublished research, pre-release reports, or confidential benchmarks
• ❌ No content obtained through employment at or breach of agreement with any firm
• ❌ No leaked internal documents of any kind

Use of terms like "McKinsey-style" or "MBB" throughout this repo is purely descriptive — referring to a publicly understood standard of consulting practice, not a claim of affiliation or endorsement by any firm.

If you are a current or former consultant and wish to contribute, please ensure any additions follow these same standards: public sources only, no client data, no confidential firm IP.

Who This Is For

What's Inside

management-consultant-claude-skill/
├── README.md                    <- You are here
├──