strategy-mcp

Professional-grade product strategy frameworks as MCP tools.

Give any MCP-compatible AI assistant — Claude Code, Cursor, Cline — instant access to 12 structured strategy frameworks. Not templates. Not prompts. Actual tools that accept your inputs, apply the framework, show the reasoning, and return specific next steps.

Why this exists

AI tools are great at generating content. They're inconsistent at applying structured thinking.

strategy-mcp is the product management layer that's been missing from the AI toolkit. Every framework a PM reaches for — RICE scoring, Jobs-to-be-Done, competitive positioning, OKR generation — encoded as a tool your AI can use natively.

Each tool returns:

• Structured analysis with reasoning (not just a score)
• 2-5 actionable next steps (not generic advice)
• Confidence indicator (High / Medium / Low) with rationale
• Pressure-test questions to challenge the analysis

Built by Sohaib Thiab — former CPO, now building AI products in public.

Install in 60 seconds

strategy-mcp is published on PyPI — install it with a single command.

Claude Code

claude mcp add strategy-mcp -- uv run --with strategy-mcp python server.py

Cursor

Add to your Cursor MCP settings (.cursor/mcp.json):

{
  "mcpServers": {
    "strategy-mcp": {
      "command": "uv",
      "args": ["run", "--with", "strategy-mcp", "python", "server.py"]
    }
  }
}

Cline

Add to your Cline MCP settings:

{
  "mcpServers": {
    "strategy-mcp": {
      "command": "uv",
      "args": ["run", "--with", "strategy-mcp", "python", "server.py"]
    }
  }
}

Plain pip

pip install strategy-mcp

Run locally (development)

git clone https://github.com/sohaibt/strategy-mcp.git
cd strategy-mcp
uv run python server.py

The 12 Tools

Prioritization

Discovery

Positioning

Business Model

Execution

Advanced

Example: RICE Scoring

Ask your AI assistant:

"Score our new AI onboarding feature using RICE. It reaches about 5,000 users per quarter, high impact, we're 80% confident, and it'll take 2 person-months."

What you get back:

{
  "feature_name": "AI Onboarding Feature",
  "rice_score": 2000.0,
  "priority_tier": "Critical",
  "score_breakdown": "RICE = (Reach x Impact x Confidence) / Effort\n     = (5,000 x 1 x 0.8) / 2\n     = 2,000.0",
  "analysis": "**AI Onboarding Feature** scores **2,000.0** — classified as **Critical** priority.\n\n- **Reach is moderate** (5,000 users/quarter)\n- **Impact is medium** (1x) per user affected.\n- **Confidence is high** (80%)\n- **Effort is moderate** (2 person-months)",
  "next_steps": [
    "Prioritize AI Onboarding Feature in the next sprint/cycle — the score supports it.",
    "Define success metrics before building so you can validate the impact estimate post-launch.",
    "Stack-rank this against your top 5 backlog items using the same RICE framework for consistency."
  ],
  "confidence": "High",
  "confidence_rationale": "The input estimates appear data-informed (high confidence %, meaningful reach).",
  "pressure_test_questions": [
    "Is the reach estimate (5,000 users/quarter) based on actual data or a gut feeling?",
    "Would the impact really be medium? What's the evidence from user research?",
    "Are there hidden dependencies that could inflate the 2-month effort estimate?"
  ]
}

Every tool follows this same structure: analysis + next steps + confidence + pressure-test questions.

Example: Business Model Canvas Review

"Review the business model for my AI analytics startup. We target mid-market SaaS companies, our value prop is real-time anomaly detection..."

The tool assesses all 9 BMC components, rates each as Strong/Adequate/Weak/Missing, identifies critical gaps, evaluates coherence between components, and tells you exactly what to fix first.

Example: Hypothesis Builder

"I have three assumptions about my product. Turn them into testable hypotheses."

Each assumption becomes a structured hypothesis with: independent/dependent variables, success metric, success threshold, suggested test method, estimated duration, and risk assessment. Hypotheses are prioritized by risk level — test the scariest ones first.

How it works

strategy-mcp is a stateless MCP server built with FastMCP. Each tool:

1. Accepts structured inputs via MCP tool parameters
2. Applies the framework logic in Python (no external API calls)
3. Returns structured JSON that your AI assistant formats into a readable response

All analysis is done locally using your inputs + the framework logic. No data leaves your machine. No API keys required.

Architecture

Your AI Assistant (Claude/Cursor/Cline)
        │
        ▼
   MCP Protocol
        │
        ▼
  strategy-mcp server (FastMCP)
        │
        ├── tools/prioritization.py    → RICE scoring
        ├── tools/discovery.py         → Assumption map, JTBD
        ├── tools/positioning.py       → Competitive positioning
        ├── tools/business_model.py    → BMC, TAM/SAM/SOM, Pricing
        ├── tools/execution.py         → OKR generator, Initiative scoper
        ├── tools/advanced.py          → Wardley assessment, Hypothesis builder
        └── tools/governance.py        → Decision log

Development

Run tests

uv run pytest tests/ -v

All 12 tools have happy-path and edge-case tests (24 tests total).

Project structure

strategy-mcp/
├── server.py              # MCP server entry point
├── tools/
│   ├── prioritization.py  # RICE score
│   ├── discovery.py       # Assumption map, JTBD
│   ├── positioning.py     # Competitive positioning
│   ├── business_model.py  # BMC, TAM/SAM/SOM, Pricing
│   ├── execution.py       # OKR generator, Initiative scoper
│   ├── advanced.py        # Wardley assessment, Hypothesis builder
│   └── governance.py      # Decision log
├── schemas/
│   └── models.py          # Pydantic models for all inputs/outputs
├── tests/
│   └── test_tools.py      # 24 tests across all 12 tools
└── pyproject.toml

Contributing

Contributions welcome! Some ways to help:

• Add a new framework tool — open an issue with the framework name and what it should do
• Improve an existing tool — better analysis logic, smarter recommendations
• Add test cases — especially edge cases and realistic product scenarios
• Fix bugs — if a tool gives bad advice, that's a bug

Please open an issue before submitting large PRs so we can discuss the approach.

License

MIT — use it however you want.

Built by

Sohaib Thiab — Former CPO, now building AI products in public.

• Mastering Product HQ — Weekly writing on product leadership

Want connected strategy execution? GetVelocity.ai takes these frameworks further — connecting your OKRs to Jira, Linear, and ClickUp with AI-powered monitoring and real-time velocity tracking.

If strategy-mcp saves you a bad decision, it's done its job.