Rulesify

May 1, 2026 · View on GitHub

Rulesify is a skill management tool for AI agents. Discover, browse, and install pre-built skills from curated GitHub repositories to enhance your AI assistant's capabilities.

What are Skills?

Skills are pre-built workflows, methodologies, and instructions that enhance AI agent capabilities. Each skill is a self-contained module that guides AI assistants through complex tasks like:

  • Test-driven development - Write tests before implementation
  • Systematic debugging - Investigate bugs with scientific method
  • Deployment workflows - Deploy to Netlify, Render, Cloudflare
  • Document creation - Generate PDFs, Word docs, presentations
  • Security reviews - Threat modeling, security best practices
  • Design workflows - Figma integration, UI generation

Features

  • Skills Registry: Browse and install 50+ skills from trusted sources
  • LLM Classification: Automatic domain and tag assignment for easy discovery
  • Domain Filtering: Find skills by category (development, testing, design, etc.)
  • Tag Search: Discover skills by capability tags
  • Direct Installation: Skills are downloaded and installed automatically (full folder copy)
  • Multi-Tool Support: Install to multiple AI tools simultaneously (Claude Code, Codex, Cursor, etc.)
  • Curated Sources: Skills from Anthropic, OpenAI, and community experts
  • Auto Reconciliation: Config automatically syncs with disk state when skills are manually deleted

Skills Registry

Skill Sources

Skills are curated from trusted GitHub repositories:

  • anthropics/skills - Anthropic's official skill collection
  • openai/skills - OpenAI's curated and system skills
  • mattpocock/skills - Matt Pocock's development skills
  • MiniMax-AI/skills - MiniMax's skill collection

LLM-Based Classification

Skills are automatically classified into 10 domains:

DomainDescriptionExample Skills
planning-and-orchestrationProject planning, task orchestrationnotion-spec-to-implementation
developmentCode development, SDKs, app buildingskill-creator, mcp-builder
designUI design, visual arts, creative workfigma-*, canvas-design
documentationDocument creation, formattingdocx, pptx, pdf
dataData analysis, researchjupyter-notebook
testingTesting, debugging, error handlingplaywright, sentry
deploymentCloud deployment, hostingnetlify-deploy, render-deploy
integrationsAPI integrations, MCP serversmcp-builder, linear
collaborationTeam communication, meetingsinternal-comms, notion-meeting-intelligence
securitySecurity reviews, threat modelingsecurity-threat-model

Each skill has up to 3 tags describing specific capabilities.

Registry Format

[skills.tdd]
name = "test-driven-development"
description = "Write tests before implementation code using TDD methodology"
source_url = "https://github.com/mattpocock/skills/tree/main/tdd"
stars = 1500
context_size = 2400
domain = "development"
last_updated = "2026-04-15"
tags = ["testing", "tdd", "best-practices"]

Installation

Homebrew (macOS)

brew tap ydeng11/rulesify
brew install rulesify

Or install directly:

brew install ydeng11/rulesify/rulesify

One-liner (Linux/macOS)

curl -sSL https://github.com/ydeng11/rulesify/releases/latest/download/install.sh | bash

After installation, restart your shell or run:

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

Manual Installation

  1. Download the latest binary from GitHub Releases
  2. Install to ~/.local/bin: 
    mkdir -p ~/.local/bin
mv rulesify-<os>-<arch> ~/.local/bin/rulesify
chmod +x ~/.local/bin/rulesify

Quick Start

# List all available skills
rulesify skill list

# Filter by domain
rulesify skill list --domain development

# Search by tags
rulesify skill list --tags testing,debugging

# View skill details
rulesify skill show playwright

# Install a skill
rulesify skill add test-driven-development

Command Reference

Skills Commands

CommandDescription
rulesify initInteractive setup - select tools and skills
rulesify skill listList installed skills
rulesify skill add <skill-id>Download and install a skill (project level)
rulesify skill add <skill-id> --globalInstall a skill globally
rulesify skill remove <skill-id>Remove a skill (prompts for confirmation)
rulesify skill updateUpdate installed skills to latest versions

Global Options

  • --config <path> - Use custom configuration file
  • --verbose - Enable detailed output for debugging

Usage Examples

Discover Skills

# Browse all skills
rulesify skill list

# Find deployment skills
rulesify skill list --domain deployment

# Find debugging skills
rulesify skill list --tags debugging

# Find testing-related skills
rulesify skill list --tags testing

Install Skills

# Initialize project with interactive selection
rulesify init

# Install a skill (downloads from GitHub, creates folders)
rulesify skill add test-driven-development

# Install globally (available in all projects)
rulesify skill add test-driven-development --global

# Remove a skill (prompts for confirmation)
rulesify skill remove test-driven-development

# Update installed skills
rulesify skill update

Filter by Domain

# Development skills
rulesify skill list --domain development

# Design skills
rulesify skill list --domain design

# Security skills
rulesify skill list --domain security

Update Registry

The registry updates weekly via GitHub Actions. To manually refresh:

# Requires GitHub and OpenRouter tokens
GH_TOKEN=xxx OPENROUTER_API_KEY=xxx cargo run --bin update-registry

# Force re-classification
cargo run --bin update-registry -- --force

# Verbose mode
cargo run --bin update-registry -- --force -v

# Custom model (default: claude-3.5-haiku)
OPENROUTER_MODEL=openai/gpt-4o-mini cargo run --bin update-registry -- --force

Development

Project Structure

rulesify/
├── Cargo.toml
├── registry.toml           # Skills registry (LLM-classified)
├── src/
│   ├── main.rs
│   ├── lib.rs
│   ├── bin/
│   │   └── update-registry.rs  # Registry automation
│   ├── cli/
│   │   └── skill.rs            # Skill commands
│   ├── models/
│   │   ├── domain.rs           # 10 domains enum
│   │   ├── skill.rs            # Skill model
│   │   └── registry.rs         # Registry model
│   ├── llm/
│   │   ├── client.rs           # OpenRouter client
│   │   ├── classifier.rs       # Batch classification
│   │   └── prompt.rs           # Prompts
│   ├── registry/
│   │   ├── source.rs           # GitHub sources
│   │   ├── github.rs           # GitHub API
│   │   ├── parser.rs           # SKILL.md parser
│   │   ├── scorer.rs           # Quality scoring
│   │   └── generator.rs        # TOML generator
│   └── tui/
│       └── skill_selector.rs   # Interactive selection
├── .github/
│   └── workflows/
│       └── update-registry.yml # Weekly updates
└── docs/
    └ plans/

Testing

cargo test        # Run all 116 tests
cargo check       # Verify compilation
cargo clippy      # Lint

Test Coverage

  • 116 total tests across all modules
  • 12 reconciliation tests (skill verification, config cleanup)
  • 7 executor tests (URL parsing, install/uninstall)
  • 10 domain tests (enum parsing, validation)
  • 24 registry tests (skills catalog, GitHub API)
  • 4 TUI tests (interactive selection)

License

MIT License

Copyright (c) 2024 Rulesify Contributors