What Frameworks or Libraries Does coreyhaines31/marketingskills Use? A Zero-Dependency Analysis

The Marketing Skills repository relies on zero external frameworks or libraries, using only native Node.js 18+ capabilities, built-in modules like fs and fetch, and standard Unix shell tools for validation.

This repository contains a curated collection of marketing skills and CLI tools designed for AI agents. Unlike typical JavaScript projects that pull in dozens of npm packages, this codebase takes a deliberately lightweight approach. The architecture consists entirely of static markdown content with YAML front-matter and zero-dependency Node.js scripts.

Runtime: Native Node.js 18+ Only

The codebase targets Node.js 18 or higher, leveraging runtime features without installing a single external package.

Built-in fetch API: All HTTP requests in the CLI tools use the native fetch implementation. In tools/clis/ga4.js and similar scripts, network calls require no axios, node-fetch, or other HTTP libraries:

// From tools/clis/ga4.js - zero imports, zero dependencies
const response = await fetch(`https://analyticsadmin.googleapis.com/v1alpha/properties`, {
  headers: { 'Authorization': `Bearer ${token}` }
});

Standard library modules: File system operations rely exclusively on Node's built-in fs module. The tools read local JSON payloads and skill definitions using require('fs') with no wrapper libraries like fs-extra.

The repository root contains no package.json file, confirming that zero npm dependencies are declared or managed.

Content Architecture: Plain Markdown and YAML

Skills are stored as static markdown files with YAML front-matter (---\nname: …\ndescription: …\n---). The repository processes these as raw text rather than using parsing libraries.

In skills/README.md and individual skill files, content is structured for direct consumption by AI agents. When an agent loads a skill—for example, via /page-cro in a Claude Code session—it reads the markdown file directly without transformation pipelines like remark, gray-matter, or markdown-it.

The registry in tools/REGISTRY.md maintains a master list of available tools as a markdown table. Accessing this programmatically requires only basic string manipulation:

// Pure Node.js - no YAML parsers or markdown libraries needed
const fs = require('fs');
const registry = fs.readFileSync('tools/REGISTRY.md', 'utf8');
// Simple string parsing replaces heavy dependencies
console.log(registry);

CLI Tools: Zero-Dependency Command-Line Scripts

The tools/clis/ directory contains standalone JavaScript files that execute without npm installation or node_modules folders.

Each script follows the same pattern demonstrated in tools/clis/wistia.js:

• Uses process.argv for argument parsing (no commander or yargs)
• Uses console for output formatting (no chalk or colors)
• Uses native fetch for API calls (no request libraries)

Running a CLI tool requires only the Node runtime:

# Execute directly - no npm install, no dependencies to download

node tools/clis/ga4.js properties list --dry-run

Validation Pipeline: Standard Unix Tools

Quality assurance bypasses the JavaScript ecosystem entirely. The validate-skills.sh script performs static validation using standard Unix utilities rather than ESLint, Prettier, or other Node-based linters.

This shell script checks front-matter constraints and markdown structure using:

• bash for scripting logic
• node --check for basic JavaScript syntax validation
• Standard text-processing utilities (grep, awk, sed)

This approach ensures the validation pipeline itself introduces no additional dependencies.

Summary

• Zero npm packages: The repository contains no package.json and no node_modules directory.
• Node.js 18+ built-ins: CLI tools rely exclusively on native fetch, fs, and process modules.
• Raw text processing: Markdown and YAML are handled as static files without parsing libraries.
• Unix-native validation: Shell scripts in validate-skills.sh enforce code quality without JavaScript tooling.
• Direct execution: Scripts run immediately with node without installation steps.

Frequently Asked Questions

Does the Marketing Skills repository use React or frontend frameworks?

No. The repository contains no React, Vue, Angular, or other frontend frameworks. It is not a web application but a collection of static markdown content and command-line utilities designed for AI agents and terminal usage.

Why is there no package.json file in the repository?

The absence of package.json is intentional and confirms the zero-dependency architecture. By omitting this file, the project signals that it requires no npm installation, no dependency resolution, and no build steps. Users clone the repository and run scripts immediately using the Node.js runtime alone.

How does the repository parse YAML front-matter without a library?

The repository does not parse YAML front-matter programmatically. Skill files in the skills/ directory use front-matter for human readability and AI context, but the content is read as raw text. AI agents and shell scripts access the markdown directly without converting YAML into data structures, eliminating the need for parsers like js-yaml.

Can I use these CLI tools if I only have Node.js installed?

Yes. As long as you have Node.js 18 or higher installed, the CLI tools in tools/clis/ function immediately. Clone the repository and run any script directly (e.g., node tools/clis/ga4.js). No npm install command is required because the scripts import only Node.js built-in modules.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}