Mckinsey Strategy Team — Design skill for Claude Code

mckinsey-strategy-team

A Claude Code **skill** that orchestrates a live *agent team* for strategic decisions. A team lead runs intake, classifies the question, then runs the team in waves on McKinsey-style frameworks: diagnose + map the market in parallel, generate **competing options from opposing mandates** and filter them to a shortlist, synthesize a recommendation, then turn a **panel of verifiers** loose on its load-bearing assumptions (a weakness ≥2 of them agree on doesn't survive) before it ships. Output: a board-ready **decision memo + narrative** that survives a leadership or board meeting.

Use it to: prepare a leadership/board session, structure a merger/M&A question, or stress-test / war-game an existing strategy.

What's inside

mckinsey-strategy-team/
├── SKILL.md        # the team-lead recipe (the orchestration)
├── README.md       # this file
├── WHY.txt         # the design rationale (what, how, value)
└── references/     # 21 McKinsey-style frameworks, 6 domains (method docs)

The 21 frameworks come from github.com/aapersh/strategy-skills-for-claude. They are not installed as separate skills — the teammates **read them as files**. That sidesteps the limitation that a teammate does not inherit a subagent's skill frontmatter.

Credits

The framework reference docs in `references/` are based on **Oria AI's "21 Strategy Skills for Claude"** — [oria.one/resources/21-strategy-skills-for-claude](https://www.oria.one/resources/21-strategy-skills-for-claude). This project adapts that material into a live agent-team orchestration; all the underlying McKinsey-style methodology and framework write-ups originate from their work. Huge thanks to the Oria team for making it available — bedankt! 🙏

Prerequisites

• Claude Code ≥ 2.1.32 (claude --version)
• Agent teams enabled. Quickest way — run this in your terminal before you start:

  export CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1

  That only applies to the Claude Code