clawfu-skills
172 expert marketing skills for AI agents — ClawFu MCP Server
claude mcp add --transport stdio guia-matthieu-clawfu-skills npx @clawfu/mcp-skills
How to use
ClawFu Skills offers 159 expert marketing methodologies for AI, packaged as structured AI instructions that you can run and apply via an MCP server. The MCP setup lets you load and execute the skills through the npx-based entry point, enabling Claude-like tooling for a broad range of marketing disciplines such as copywriting, positioning, persuasion, and negotiation. You can also run the MCP server by adding a clawfu entry to your .mcp.json, which makes it easy to switch between multiple MCPs in your workspace and orchestrate multi-skill workflows.
To use the tools, start the server with npx @clawfu/mcp-skills (or add the clawfu entry to your MCP configuration). Once loaded, you can browse and download individual skills from clawfu.com/skills and apply them to your AI tasks. The skills are organized across categories like Content, Strategy, Sales, Growth, and more, each backed by expert methodologies and the "What Claude Does vs What You Decide" guidance to keep a clear human/AI boundary in your prompts. You can mix and match skills to build multi-skill workflows that cover ideation, drafting, validation, and optimization in one session.
How to install
Prerequisites:
- Node.js and npm installed on your system (npx is included with npm).
- Internet access to fetch the MCP package from npm.
Installation steps:
-
Install and run the MCP server via npx:
- Command: npx @clawfu/mcp-skills
-
Optional: register the MCP in your .mcp.json for automatic startup:
- Add the following snippet to your .mcp.json:
{ "mcpServers": { "clawfu": { "command": "npx", "args": ["@clawfu/mcp-skills"] } } }
-
If you prefer to clone the repository first and run locally, clone then install dependencies and start as appropriate for your setup (the package on npm is the recommended entry point):
- git clone https://github.com/guia-matthieu/clawfu-skills.git
- cd clawfu-skills
- npm install
- npx @clawfu/mcp-skills
Notes:
- The MCP is designed to be invoked via npx for quick, ephemeral usage or integrated into a larger MCP workflow via .mcp.json.
- No separate server process is required beyond the Node.js/npm runtime.
Additional notes
Tips and considerations:
- The MCP package is published as @clawfu/mcp-skills; use the npx command to run without local installation."- If you manage multiple MCPs, consider adding each to your .mcp.json with distinct server names for easier orchestration.
- Browse clawfu.com/skills to discover individual skills before running them in a session; each skill includes metadata and usage notes.
- If you encounter network issues, ensure your environment can reach npm registries and that your Node.js version is compatible with the package requirements.
- When using multiple skills, plan prompts to preserve the intended division of labor per the Skill Format (e.g., including a "What Claude Does vs What You Decide" section in your prompts).
- For reproducibility, pin the package version when possible or use a dedicated environment/virtual environment to avoid version drift.
Related MCP Servers
Continuous-Claude-v3
Context management for Claude Code. Hooks maintain state via ledgers and handoffs. MCP execution without context pollution. Agent orchestration with isolated context windows.
skillz
An MCP server for loading skills (shim for non-claude clients).
skrills
Coordinate skills between Codex, Copilot, and Claude Code. Validates, analyzes, and syncs skills, subagents, commands, and configuration between multiple CLIs.
mem0 -selfhosted
Self-hosted mem0 MCP server for Claude Code. Run a complete memory server against self-hosted Qdrant + Neo4j + Ollama while using Claude as the main LLM.
stt -linux
Local speech-to-text MCP server for Tmux on Linux (for use not only with Claude Code)
cc-audit
AI-free static security scanner for Claude Code artifacts (Skills, Hooks, MCP configs). Detects data exfiltration, prompt injection, and supply chain risks with deterministic, reproducible results.