INSTALL.md

generate-tech-stack — Installation Guide

Type /generate-tech-stack in any supported tool to get a visual tech stack for your project.

The skill lives at: ~/.claude/skills/generate-tech-stack/

~/.claude/skills/generate-tech-stack/
├── SKILL.md                  ← Claude Code skill definition
├── INSTALL.md                ← this file
├── scripts/
│   └── analyze.py            ← core scanner + HTML renderer
├── mcp/
│   ├── server.py             ← MCP stdio server
│   └── requirements.txt
└── copilot/
    ├── index.js              ← GitHub Copilot Extension (Express)
    ├── package.json
    └── openai_function.json  ← Antigravity / OpenAI function def

---

Claude Code

The skill is already installed. Just use:

/generate-tech-stack

To reinstall or update:

# Clone or copy the skill files
mkdir -p ~/.claude/skills/generate-tech-stack/{scripts,mcp,copilot}

# Core analyzer
curl -L https://raw.githubusercontent.com/yourorg/generate-tech-stack/main/scripts/analyze.py \
     -o ~/.claude/skills/generate-tech-stack/scripts/analyze.py

# Skill definition
curl -L https://raw.githubusercontent.com/yourorg/generate-tech-stack/main/SKILL.md \
     -o ~/.claude/skills/generate-tech-stack/SKILL.md

---

MCP (Claude Desktop, VS Code, Cursor, Zed, Continue, Antigravity, Windsurf)

Any host that supports the Model Context Protocol can use the MCP server.

Install the dependency:

pip install mcp

Configure your MCP host — replace /home/<you> with your actual home directory:

Claude Desktop

~/Library/Application Support/Claude/claude_desktop_config.json (Mac)
%APPDATA%\Claude\claude_desktop_config.json (Windows)

{
  "mcpServers": {
    "generate-tech-stack": {
      "command": "python3",
      "args": ["/home/<you>/.claude/skills/generate-tech-stack/mcp/server.py"]
    }
  }
}

VS Code (.vscode/mcp.json or user settings)

{
  "servers": {
    "generate-tech-stack": {
      "type": "stdio",
      "command": "python3",
      "args": ["/home/<you>/.claude/skills/generate-tech-stack/mcp/server.py"]
    }
  }
}

Cursor (~/.cursor/mcp.json)

{
  "mcpServers": {
    "generate-tech-stack": {
      "command": "python3",
      "args": ["/home/<you>/.claude/skills/generate-tech-stack/mcp/server.py"]
    }
  }
}

Zed (.zed/settings.json)

{
  "context_servers": {
    "generate-tech-stack": {
      "command": {
        "path": "python3",
        "args": ["/home/<you>/.claude/skills/generate-tech-stack/mcp/server.py"]
      }
    }
  }
}

Restart your tool, then use:

/generate-tech-stack
# or:
"generate my tech stack"
"scan this project and show all tools visually"

MCP tools exposed:

• generate_tech_stack — scans project, writes TECH_STACK.html, opens in browser
• list_tech_stack — returns JSON summary, no file written

---

Antigravity / OpenAI-compatible tools

Register the function definition from copilot/openai_function.json in your assistant config. When the model calls generate_tech_stack, execute:

python3 ~/.claude/skills/generate-tech-stack/scripts/analyze.py <project_dir> <output_file>

---

GitHub Copilot Extension

The Copilot Extension is a GitHub App that you deploy as a small Express server.

Prerequisites:

• GitHub account with Copilot access
• A public HTTPS endpoint (use ngrok for local dev)

Start the server:

cd ~/.claude/skills/generate-tech-stack/copilot
npm install
npm start   # listens on port 3000

For local dev, expose it:

ngrok http 3000
# copy the https://xxxx.ngrok.io URL

Register the GitHub App:

1. GitHub → Settings → Developer settings → GitHub Apps → New GitHub App
2. Set Homepage URL and Callback URL to your ngrok/deployment URL
3. Enable Copilot Extension and set Agent URL: https://your-url/agent
4. Permissions: Copilot chat → Read & Write
5. Install the app on your account or organisation

Usage (in GitHub Copilot Chat):

@generate-tech-stack /generate-tech-stack
@generate-tech-stack /generate-tech-stack /path/to/project

---

Quick test (any platform)

Run the analyzer directly to verify it works:

python3 ~/.claude/skills/generate-tech-stack/scripts/analyze.py . ~/TECH_STACK.html
# then open ~/TECH_STACK.html in your browser

---

What gets detected

Source file	Detected tools
requirements.txt / pyproject.toml	Python packages
package.json	Node.js / npm packages
go.mod	Go language
Cargo.toml	Rust language
pom.xml / build.gradle	Java / Kotlin
Gemfile	Ruby language
composer.json	PHP language
Dockerfile	Docker
docker-compose.yml	Docker Compose
.github/workflows/	GitHub Actions
.gitlab-ci.yml	GitLab CI
alembic.ini	Alembic migrations
nginx.conf / Caddyfile	Reverse proxy
tsconfig.json / src/**/*.ts	TypeScript

Output sections

Every generated TECH_STACK.html includes:

1. Stat row — Total Tools · Categories · AI Backends · Data Stores
2. Architecture diagram — layered flow adapted to what's detected
3. Bar chart — horizontal bars per category, colour-matched
4. Tool cards — one card per category, each tool with dot · name · desc · badge