Skip to content

alirezarezvani/claude-code-skill-factory

BranchesTags

Repository files navigation

Claude Code Skills & Agents Factory

A comprehensive toolkit for generating production-ready Claude Skills and Claude Code Agents at scale. This repository provides templates, examples, and powerful prompt engineering systems to create custom skills and specialized agents for Claude AI across all platforms (Claude apps, Claude Code, and API).

Table of Contents

What Are Claude Skills?

Claude Skills are specialized capabilities that teach Claude how to perform specific tasks. They're packaged as folders containing:

  • SKILL.md: Structured instructions with YAML frontmatter
  • Python files (optional): Functional code for calculations, data processing, or file generation
  • Sample data: JSON examples showing inputs and expected outputs
  • Usage guide: Clear invocation examples

Skills are composable (work together), portable (same format everywhere), and efficient (loaded only when relevant).

What Are Claude Code Agents?

Claude Code Agents (also called sub-agents) are specialized AI assistants that handle specific types of tasks. They're single Markdown files with YAML frontmatter containing:

  • Enhanced YAML frontmatter: name, description, tools, model, color, field, expertise, MCP integrations
  • System prompt: Detailed instructions for the agent's behavior and approach
  • Auto-invocation: Claude automatically uses them when the description matches the task

Agents are focused (one responsibility), efficient (separate context window), flexible (configurable tool access), and composable (multiple agents work together on complex workflows).

Key Difference:

  • Skills = Multi-file capabilities (folders with SKILL.md + Python + samples)
  • Agents = Single-file specialists (.md files in .claude/agents/)

Repository Contents

claude-code-skills-factory/
├── README.md                              # This file
├── CLAUDE.md                              # Repository guidance for Claude Code
├── claude-skills-instructions.md          # Full Skills documentation from Anthropic
├── claude-agents-instructions.md          # Full Agents documentation from Anthropic
├── claude-skills-examples/                # Reference skill implementations
│   ├── analyzing_financial_statements.md
│   ├── calculate_ratios.py
│   ├── interpret_ratios.py
│   ├── creating-financial-models.md
│   ├── dcf_model.py
│   ├── sensitivity_analysis.py
│   ├── brand_guidelines.md
│   └── apply_brand.py
├── generated-skills/                      # Production-ready generated skills
│   ├── aws-solution-architect/            # AWS architecture (53 KB)
│   ├── content-trend-researcher/          # Multi-platform content research (35 KB)
│   ├── ms365-tenant-manager/              # Microsoft 365 administration (40 KB)
│   ├── psychology-advisor/                # Mental wellness & CBT techniques (31 KB)
│   ├── agent-factory/                     # Claude Code agent generation (12 KB)
│   └── prompt-factory/                    # Prompt generation powerhouse (427 KB)
└── documentation/
    └── templates/
        ├── SKILLS_FACTORY_PROMPT.md       # Template for generating Skills
        └── AGENTS_FACTORY_PROMPT.md       # Template for generating Agents

Quick Start

Option A: Generate Claude Skills (Multi-file Capabilities)

1. Use the Skills Factory Prompt

Open documentation/templates/SKILLS_FACTORY_PROMPT.md and scroll to the bottom.

Fill in the template variables:

=== FILL IN YOUR DETAILS BELOW ===

BUSINESS_TYPE: SaaS startup (project management tool)

USE_CASES: Analyze user feedback sentiment, Generate feature prioritization reports, Create customer success playbooks

NUMBER_OF_SKILLS: 3

ADDITIONAL_CONTEXT: Use data-driven decision making, modern tech stack

2. Generate Your Skills

Copy the entire prompt (including your filled variables) and paste it into:

  • Claude.ai (any plan with Skills enabled)
  • Claude Code
  • Claude API (with Code Execution Tool)

Claude will generate complete skill packages with:

  • ✅ Properly formatted SKILL.md files (kebab-case names in YAML)
  • ✅ Python implementation files (when functional code is needed)
  • ✅ Sample input/output JSON files
  • ✅ HOW_TO_USE.md with invocation examples
  • ✅ ZIP files ready for import into Claude AI desktop

3. Install Your Skills

Claude AI Desktop:

  • Import the generated .zip files directly through the Skills menu

Claude Code:

  • Copy skill folders to ~/.claude/skills/

Claude Apps (Browser):

  • Use the "skill-creator" skill to import

API:

  • Use the /v1/skills endpoint to upload

Option B: Generate Claude Code Agents (Single-file Specialists)

1. Use the Agents Factory Prompt

Open documentation/templates/AGENTS_FACTORY_PROMPT.md and scroll to the bottom.

Fill in the template variables:

=== FILL IN YOUR DETAILS BELOW ===

AGENT_NAME: api-integration-specialist
AGENT_TYPE: Implementation
DOMAIN_FIELD: backend
DESCRIPTION: API integration expert. Use when building API clients.
TOOLS_NEEDED: Read, Write, Edit, Bash
MODEL: sonnet
COLOR: green
EXPERTISE_LEVEL: expert
MCP_TOOLS: mcp__github

2. Generate Your Agent

Copy the entire prompt and paste into Claude. You'll get a complete agent .md file:

---
name: api-integration-specialist
description: API integration expert...
tools: Read, Write, Edit, Bash
model: sonnet
color: green
field: backend
expertise: expert
mcp_tools: mcp__github
---

System prompt with detailed instructions...

3. Install Your Agent

Project-level (shared with team):

cp api-integration-specialist.md .claude/agents/

User-level (available everywhere):

cp api-integration-specialist.md ~/.claude/agents/

Agent auto-invokes when Claude detects relevant tasks!

Example Skills Included

This repository includes three fully-functional example skills:

1. Analyzing Financial Statements

Purpose: Calculate and interpret financial ratios (ROE, ROA, liquidity, leverage, etc.)

Files:

Pattern: Calculation engine + interpretation layer

2. Creating Financial Models

Purpose: DCF valuation, Monte Carlo simulation, sensitivity analysis

Files:

Pattern: Historical data → projections → valuation

3. Applying Brand Guidelines

Purpose: Apply consistent corporate branding to documents

Files:

Pattern: Brand definition + application logic

Generated Skills

This repository also includes production-ready skills generated using the Skills Factory Prompt, demonstrating the quality and completeness of output you can expect.

4. AWS Solution Architect

Purpose: Expert AWS architecture design for startups - serverless, scalable, and cost-effective cloud infrastructure

Files:

Capabilities:

  • Serverless architecture design (Lambda, API Gateway, DynamoDB, EventBridge)
  • Infrastructure as Code (CloudFormation, CDK, Terraform)
  • Cost optimization and budget management
  • Security best practices and compliance
  • CI/CD pipeline setup
  • Multi-region deployment strategies

Pattern: Architecture design → IaC templates → cost optimization

5. Content Trend Researcher

Purpose: Advanced content research analyzing trends across 10+ platforms to generate data-driven article outlines

Files:

Capabilities:

  • Multi-platform trend analysis (Google Trends, Reddit, YouTube, Medium, Substack, LinkedIn, X, etc.)
  • User intent analysis (informational, commercial, transactional)
  • Content gap identification
  • SEO-optimized article outline generation
  • Platform-specific best practices and publishing strategies

Pattern: Trend analysis → intent analysis → content gap discovery → outline generation

6. Microsoft 365 Tenant Manager

Purpose: Comprehensive M365 tenant administration for setup, security, user management, and organizational optimization

Files:

Capabilities:

  • Tenant setup and configuration
  • User and group lifecycle management
  • Security and compliance policies (Conditional Access, MFA, DLP)
  • SharePoint, OneDrive, and Teams administration
  • Exchange Online management
  • PowerShell automation script generation
  • License management and cost optimization

Pattern: Configuration requirements → PowerShell scripts → validation checklists

7. Prompt Factory

Purpose: World-class prompt powerhouse for generating production-ready mega-prompts for any role, industry, and task

Files:

Capabilities:

  • 69 comprehensive presets across 15 professional domains
  • Intelligent 7-question flow for requirement gathering
  • Multiple output formats (XML/Claude/ChatGPT/Gemini)
  • 7-point quality validation before delivery
  • Contextual best practices from OpenAI, Anthropic, Google
  • Core & Advanced modes with testing scenarios
  • Complete coverage: Technical (8), Business (8), Legal (4), Finance (4), HR (4), Design (4), Customer (4), Executive (7), Specialized-Technical (6), Research (3), Creative-Media (4), Manufacturing (4), R&D (2), Regulatory (1), Specialized (3)

Pattern: 7-question flow → preset selection → quality validation → multi-format output

Key Features

🎯 Production-Ready Output

Every generated skill follows enterprise standards:

  • Proper YAML frontmatter with kebab-case names
  • Type-annotated Python with error handling
  • Safe operations (division by zero, missing data)
  • Minimal, focused sample data

🔗 Composable Design

Skills are designed to work together:

  • data-extractor → data-analyzer → report-generator → brand-formatter
  • No duplicate functionality
  • Clear input/output contracts

📦 Complete Packaging

Each skill includes everything needed:

  • SKILL.md with structured documentation
  • Python files (multi-file when needed)
  • sample_input.json and expected_output.json
  • HOW_TO_USE.md with invocation examples
  • .zip file containing entire skill folder

⚡ Smart Implementation Detection

The prompt automatically determines:

  • When Python code is needed vs. prompt-only
  • Multi-file structure for complex skills
  • Appropriate sample data formats

Skill Architecture Patterns

Prompt-Only Skills

For instructional or template-based tasks:

  • Style guides and tone of voice
  • Decision frameworks
  • Content guidelines
  • Brand voice standards

No Python needed - just structured instructions in SKILL.md

Functional Skills

For computation, processing, or generation:

  • Financial calculations
  • Data analysis and transformation
  • File generation (Excel, PDF, etc.)
  • API interactions

Python files included - class-based, type-annotated, with safe operations

Multi-File Skills

For complex functionality:

  • calculate_*.py - Core computations
  • interpret_*.py - Analysis and insights
  • format_*.py - Output formatting
  • validate_*.py - Input validation

Best Practices

When Creating Skills

  1. Keep skills focused - One clear purpose per skill
  2. Make them composable - Output of one skill feeds another
  3. Validate inputs - Handle missing or invalid data gracefully
  4. Use kebab-case - For YAML name fields and folder names
  5. Provide context - Industry-specific considerations in documentation
  6. Include examples - Clear invocation patterns in HOW_TO_USE.md

When Using Skills

  1. Match to task - Use the skill that fits your specific need
  2. Provide complete data - Skills work best with full input data
  3. Check outputs - Validate results against your requirements
  4. Combine skills - Chain multiple skills for complex workflows

Customization Guide

Adapting Example Skills

To customize the included examples:

  1. Update YAML frontmatter - Change name and description to match your use case
  2. Modify constants - Replace Acme Corporation branding with yours
  3. Adjust calculations - Modify formulas for your industry
  4. Extend functionality - Add new methods to existing classes
  5. Simplify - Remove features you don't need

Creating New Skills

Use the Skills Factory Prompt with your specific:

  • Business type and industry
  • Specific use cases and tasks
  • Number of skills needed
  • Technical requirements and constraints

YAML Frontmatter Rules

CRITICAL: Every SKILL.md must start with properly formatted YAML:

---
name: skill-name-in-kebab-case
description: Brief one-line description of what this skill does
---

Correct:

---
name: financial-ratios
description: Calculates key financial ratios from financial statement data
---

Incorrect:

---
name: Financial Ratios  ❌ Title Case
name: financial_ratios  ❌ snake_case
name: financialRatios   ❌ camelCase
---

Technical Details

Python Standards

All Python files follow these conventions:

  • Type hints for all functions and classes
  • Docstrings with Args and Returns sections
  • Safe operations (e.g., safe_divide to avoid division by zero)
  • Class-based structure for stateful operations
  • Modular design with separation of concerns

File Naming

  • Folders: skill-name/ (kebab-case)
  • YAML name: name: skill-name (kebab-case)
  • Python files: calculate_metrics.py (snake_case - Python convention)
  • ZIP files: skill-name.zip (kebab-case)

Sample Data

Keep sample files minimal and focused:

  • sample_input.json: Just enough data to test the skill
  • expected_output.json: Clear structure showing what skill produces
  • Use realistic but simple values
  • Include all required fields

Resources

Documentation

Support

Real-World Examples

The generated-skills/ folder contains actual production-ready skills created using the Skills Factory Prompt:

AWS Solution Architect

Generated for startup infrastructure needs - includes architecture design, cost optimization, and IaC templates.

See: generated-skills/aws-solution-architect/

Content Trend Researcher

Generated for content creators and marketers - analyzes trends across 10+ platforms and creates data-driven outlines.

See: generated-skills/content-trend-researcher/

Microsoft 365 Tenant Manager (40 KB)

Generated for IT administrators - manages M365 tenant setup, security, and generates PowerShell automation scripts.

See: generated-skills/ms365-tenant-manager/

Psychology Advisor (31 KB)

Evidence-based mental wellness skill with CBT techniques, mindfulness exercises, stress management, and emotional regulation tools.

Key Features: 10 cognitive distortion types, 4 breathing techniques, RAIN method, stress assessment, behavioral activation

See: generated-skills/psychology-advisor/

Content Trend Researcher (35 KB)

Multi-platform content research analyzing trends across Google, Reddit, YouTube, Medium, LinkedIn, X, Substack, and more to generate data-driven article outlines.

Key Features: User intent analysis, content gap discovery, platform-specific strategies, SEO-optimized outlines

See: generated-skills/content-trend-researcher/

Agent Factory (12 KB)

Claude Code agent generation system that creates custom agents/sub-agents with enhanced YAML frontmatter, tool patterns, and MCP integration.

Key Features: Enhanced YAML (color, field, expertise), tool access patterns, execution safety, MCP tool suggestions

See: generated-skills/agent-factory/ Template: documentation/templates/AGENTS_FACTORY_PROMPT.md

Prompt Factory (427 KB)

World-class prompt powerhouse that generates production-ready mega-prompts for any role, industry, and task through intelligent 7-question flow.

Key Features: 69 comprehensive presets across 15 professional domains (Technical, Business, Legal, Finance, HR, Design, Customer, Executive, Manufacturing, R&D, Regulatory, Specialized-Technical, Research, Creative-Media), multiple output formats (XML/Claude/ChatGPT/Gemini), 7-point quality validation gates, contextual best practices from OpenAI/Anthropic/Google

Domains Covered: Full-Stack Engineer, DevOps, Product Manager, Legal Counsel, CFO, HR Manager, UI/UX Designer, Customer Success, CEO, ML Engineer, Research Scientist, Copywriter, Manufacturing Engineer, Clinical Specialist, and 55+ more roles

See: generated-skills/prompt-factory/

Each includes full implementation, sample data, HOW_TO_USE guide, and ready-to-import ZIP files.

Contributing

This is a reference repository. To contribute:

  1. Fork the repository
  2. Add new example skills to claude-skills-examples/
  3. Ensure skills follow all formatting standards
  4. Include complete implementation with samples
  5. Submit a pull request

License

This repository provides examples and templates for creating Claude Skills. The skills you generate using these templates are yours to use as you see fit.

Version

Current Version: 1.2.0 Last Updated: October 23, 2025 Compatible With: Claude Skills (all platforms) and Claude Code Agents

Latest Changes (v1.2.0):

  • Added Prompt Factory skill (427 KB) - World-class prompt generation with 69 presets across 15 domains
  • Expanded coverage: Technical, Business, Legal, Finance, HR, Design, Customer, Executive, Manufacturing, R&D, Regulatory, Specialized-Technical, Research, Creative-Media domains
  • Multi-format output support (XML/Claude/ChatGPT/Gemini)

Quick Reference

Create Skills: Use SKILLS_FACTORY_PROMPT.md Create Agents: Use AGENTS_FACTORY_PROMPT.md Create Prompt Builders: Use PROMPTS_FACTORY_PROMPT.md See Examples: Check claude-skills-examples/ Generated Skills: Explore generated-skills/ Read Guide: See CLAUDE.md for repository structure Learn More (Skills): Read claude-skills-instructions.md Learn More (Agents): Read claude-agents-instructions.md

Ready to build? Open a prompt template, fill in your details, and start generating production-ready skills or agents!

About

A comprehensive toolkit for generating production-ready Claude Skills and Claude Code Agents at scale.

medium.com/@alirezarezvani/why-agent-skills-will-transform-how-we-build-ai-32daee24fc8a

Topics

ai-agents ai-tools claude-ai claude-code claude-skills claude-skills-creator

Resources

Readme
Activity

Stars

10 stars

Watchers

0 watching

Forks

2 forks
Report repository

Releases

No releases published

Packages

No packages published

Contributors 2

Languages