skillhub 0.1.0

Package manager for AI agent workflows

SkillHub CLI

Command-line interface for the SkillHub registry of AI agent workflows

What is SkillHub CLI?

SkillHub CLI is a package manager for AI agent workflows. It allows you to search, download, and manage reusable AI agent task sequences from the SkillHub Registry.

Think of it like:

• Homebrew for AI agent workflows
• npm for coding tasks
• pip for automation recipes

Architecture

Installation

From PyPI (Coming Soon)

pip install skillhub

From Source

git clone https://github.com/v1k22/skillhub-cli.git
cd skillhub-cli
pip install -e .

Quick Start

# Search for skills
skillhub search "react"

# List all available skills
skillhub list

# Download a skill
skillhub pull react-app-setup

# Show skill details
skillhub show benchmark-qwen

# Validate a skill file
skillhub validate my-skill.md

# Show configuration and stats
skillhub info

Commands

skillhub search <query>

Search for skills by name, description, tags, or category.

# Search for React-related skills
skillhub search "react"

# Search for benchmarking skills
skillhub search "benchmark"

# Limit results
skillhub search "python" --limit 5

# Filter by category
skillhub search "model" --category ml-ops

skillhub list

List all available skills in the registry.

# List all skills
skillhub list

# Filter by category
skillhub list --category web-development

# Refresh index from registry
skillhub list --refresh

skillhub pull <skill-name>

Download a skill to your local directory.

# Pull to current directory
skillhub pull benchmark-qwen

# Pull to specific directory
skillhub pull react-app-setup --output ~/skills

skillhub show <skill-name>

Show detailed information about a skill.

skillhub show model-deployment

skillhub validate <file>

Validate a skill file format (useful when creating new skills).

skillhub validate my-new-skill.md

skillhub info

Show SkillHub configuration and registry statistics.

skillhub info

Configuration

SkillHub stores its configuration in ~/.skillhub/config.json.

Default configuration:

{
  "registry_url": "https://github.com/v1k22/skillhub-registry",
  "registry_raw_url": "https://raw.githubusercontent.com/v1k22/skillhub-registry/main",
  "cache_dir": "~/.skillhub/cache",
  "skills_dir": "~/.skillhub/skills",
  "index_url": "https://raw.githubusercontent.com/v1k22/skillhub-registry/main/index.json",
  "cache_ttl": 3600
}

You can manually edit this file to customize URLs or cache settings.

Examples

Finding and Using a Skill

# 1. Search for what you need
$ skillhub search "benchmark model"

Found 2 skill(s)
┌─────────────────┬──────────────┬─────────────────────────────┬────────────────┐
│ Name            │ Category     │ Description                 │ Tags           │
├─────────────────┼──────────────┼─────────────────────────────┼────────────────┤
│ benchmark-qwen  │ benchmarking │ Benchmark Qwen 3B model...  │ llm, benchmark │
└─────────────────┴──────────────┴─────────────────────────────┴────────────────┘

# 2. Get more details
$ skillhub show benchmark-qwen

# 3. Pull the skill
$ skillhub pull benchmark-qwen

✓ Skill downloaded successfully!
Location: ./benchmark-qwen.md

# 4. Follow the steps in the file
$ cat benchmark-qwen.md

Listing Skills by Category

$ skillhub list --category ml-ops

📚 Available Skills: 1 total

ML-OPS (1 skills)
  model-deployment v1.0.0
    Deploy ML model as REST API using FastAPI with Docker...

Creating and Validating a New Skill

# Create a skill using the template
$ curl -O https://raw.githubusercontent.com/v1k22/skillhub-registry/main/templates/skill-template.md

# Edit the template
$ vim skill-template.md

# Validate it
$ skillhub validate skill-template.md

✅ Perfect! No issues found.

Use Cases

1. Reproducible Development Environments

# Set up a React development environment
skillhub pull react-app-setup
cat react-app-setup.md  # Follow the steps

2. Model Benchmarking

# Benchmark a model on different hardware
skillhub pull benchmark-qwen
# Run on machine A, save results
# Run on machine B, compare

3. Team Onboarding

# New developer joining the team
skillhub pull etl-pipeline
# Everyone uses the same setup process

4. Learning Best Practices

# Learn how to deploy ML models
skillhub pull model-deployment
# See production-ready setup

Features

• 🔍 Search: Find skills by keywords, tags, or categories
• 📥 Download: Pull skills to local directory for use
• ✅ Validate: Check skill files for proper format
• 📊 Stats: View registry statistics and info
• 💾 Cache: Local caching for fast repeated access
• 🎨 Beautiful Output: Rich terminal formatting with colors and tables

Development

Running from Source

# Clone the repository
git clone https://github.com/v1k22/skillhub-cli.git
cd skillhub-cli

# Install in development mode
pip install -e .

# Run CLI
skillhub --help

Running Tests

# Install test dependencies
pip install pytest pytest-cov

# Run tests
pytest

# Run with coverage
pytest --cov=skillhub

Contributing

We welcome contributions! Please see the main SkillHub Registry CONTRIBUTING.md for guidelines.

Reporting Bugs

Open an issue on GitHub Issues with:

• Description of the bug
• Steps to reproduce
• Expected vs actual behavior
• CLI version (skillhub --version)
• OS and Python version

Roadmap

• Basic CLI commands (search, list, pull)
• Skill validation
• Local caching
• Vector search for semantic matching
• Skill execution engine (skillhub run)
• Shell completions (bash, zsh, fish)
• Offline mode
• Skill updates notification
• Interactive mode

Troubleshooting

Command not found after installation

Make sure your Python scripts directory is in PATH:

export PATH="$PATH:$HOME/.local/bin"

SSL Certificate errors

If you get SSL errors:

pip install --upgrade certifi

Cache issues

Clear the cache:

rm -rf ~/.skillhub/cache

License

MIT License - see LICENSE file for details.

Links

• SkillHub Registry - The skill repository
• PyPI Package - Install via pip
• GitHub Issues - Report bugs
• Documentation - Full documentation (coming soon)

---

Made with ❤️ by the SkillHub community