Skip to content

installation.md

Latest commit

 

History

History
218 lines (158 loc) · 4.91 KB

installation.md

File metadata and controls

218 lines (158 loc) · 4.91 KB
Raw

Installation and Setup

🌐 Language / Sprache: English | Deutsch

Overview

The 5-Step Algorithm Skill can be integrated into Claude in several ways. This guide covers all common scenarios.

Option 1: Claude.ai (Web, Desktop, Mobile)

Prerequisites

  • Claude.ai account (free or Pro/Team)
  • Access to the Skills feature

Step-by-Step Guide

  1. Download the skill file

    # Via browser: Download SKILL.md from GitHub
# Via terminal:
curl -O https://raw.githubusercontent.com/malkreide/musk-algorithm-skill/main/SKILL.md

  2. Import into Claude.ai

    • Open claude.ai
    • Click on your profile picture (bottom left)
    • Select Settings
    • Navigate to Skills
    • Click + Add Skill
    • Select Upload from file
    • Choose the downloaded SKILL.md file
    • Click Save

  3. Activate the skill

    • The skill is now in your skill library
    • Activate it globally or only for specific projects

Verification

Test the skill with:

Can you explain the Musk Algorithm to me?

Claude should explain the 5 steps sequentially.

Option 2: Claude API (Programmatic)

Prerequisites

  • Anthropic API Key
  • Python 3.8+ or Node.js 16+

Python Example

import anthropic

# Load skill content
with open("SKILL.md", "r") as f:
    skill_content = f.read()

client = anthropic.Anthropic(api_key="your-api-key")

message = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=4096,
    system=skill_content,  # Skill as system prompt
    messages=[
        {
            "role": "user",
            "content": "Analyze the procurement process."
        }
    ]
)

print(message.content)

Node.js Example

import Anthropic from "@anthropic-ai/sdk";
import fs from "fs";

// Load skill content
const skillContent = fs.readFileSync("SKILL.md", "utf8");

const client = new Anthropic({
  apiKey: process.env.ANTHROPIC_API_KEY,
});

const message = await client.messages.create({
  model: "claude-sonnet-4-20250514",
  max_tokens: 4096,
  system: skillContent,
  messages: [
    {
      role: "user",
      content: "Analyze the procurement process.",
    },
  ],
});

console.log(message.content);

Option 3: MCP Server (Model Context Protocol)

Prerequisites

  • MCP server setup (see MCP Docs)
  • Claude Desktop App or compatible tool

Step 1: Create MCP Resource

Create a mcp_server.py:

from mcp.server import Server
from mcp.types import Resource
import asyncio

server = Server("musk-algorithm-skill")

@server.list_resources()
async def list_resources() -> list[Resource]:
    return [
        Resource(
            uri="skill://musk-algorithm",
            name="5-Step Algorithm",
            mimeType="text/markdown",
            description="5-step process analysis for public administration"
        )
    ]

@server.read_resource()
async def read_resource(uri: str) -> str:
    if uri == "skill://musk-algorithm":
        with open("SKILL.md", "r") as f:
            return f.read()
    raise ValueError(f"Unknown resource: {uri}")

async def main():
    async with server.start():
        await asyncio.Event().wait()

if __name__ == "__main__":
    asyncio.run(main())

Step 2: Register MCP Server in Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "musk-algorithm": {
      "command": "python",
      "args": ["/path/to/mcp_server.py"]
    }
  }
}

Step 3: Restart Claude Desktop

The skill is now available as a resource.

Option 4: Local File Reference (Quick Test)

For quick testing without permanent installation:

  1. Download SKILL.md
  2. In Claude.ai: Upload the file to a chat
  3. Write: "Use the methodology from this file for the following analysis: [your process]"

Drawback: Works only for the current chat, not persistent.

Troubleshooting

"Skill not found"

  • Ensure the file was uploaded correctly
  • Check whether the skill is activated in settings

"Skill doesn't work as expected"

  • Ensure you're using the latest version of SKILL.md
  • Check whether the skill content is complete (~15 KB file size)

API Integration: "System prompt too long"

  • The skill is approximately 15 KB — use a model with a large context window (e.g., Claude Sonnet 4)
  • Alternatively: Compress the skill to its core steps

Best Practices

  1. Project-specific activation: Only activate the skill in projects where process optimization is relevant
  2. Combination with other skills: The 5-Step Algorithm works well combined with:
    • Document creation skills (for reports)
    • Data analysis skills (for metrics)
  3. Regular updates: Check monthly for new versions in the GitHub repo

Support

For questions or issues: