Installation Guide
Comprehensive installation instructions for Skill Engine on all supported platforms.
System Requirements
Minimum Requirements
- OS: Linux, macOS 11+, or Windows 10+ (with WSL2)
- RAM: 512MB available memory
- Disk: 50MB for binary + storage for skills
No Rust compiler needed! Skill Engine is distributed as a pre-compiled binary.
Optional Requirements
- Docker: For running Docker-based skills (containerized tools)
- Rust: Only needed if building from source (development)
Installation Methods
Method 1: One-Liner Install (Recommended)
The easiest way to install Skill Engine:
curl -fsSL https://dqkbk9o7ynwhxfjx.public.blob.vercel-storage.com/install.sh | shThis script:
- Detects your OS and architecture automatically
- Downloads the pre-compiled binary (no Rust needed!)
- Installs to
~/.skill-engine/bin/skill - Adds to your PATH (with permission)
Verify installation:
skill --version
# Output: skill 0.3.0Install specific version:
curl -fsSL https://dqkbk9o7ynwhxfjx.public.blob.vercel-storage.com/install.sh | SKILL_VERSION=0.2.0 shNon-interactive install (for CI/CD):
# Won't modify PATH, won't ask for confirmation
SKILL_NO_MODIFY_PATH=1 curl -fsSL https://dqkbk9o7ynwhxfjx.public.blob.vercel-storage.com/install.sh | shCustom install directory:
SKILL_INSTALL_DIR=/opt/skill curl -fsSL https://dqkbk9o7ynwhxfjx.public.blob.vercel-storage.com/install.sh | shMethod 2: Build from Source (Optional)
For developers who want to build from source or contribute:
Requirements: Rust 1.75.0 or later (rustup.rs)
# Install Rust (if not already installed)
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
# Clone repository
git clone https://github.com/kubiyabot/skill.git
cd skill
# Build and install
cargo install --path crates/skill-cliPlatform-Specific Notes
Linux
The one-liner install works on all Linux distributions. No additional setup needed!
Optional: Install Docker (for Docker-based skills):
Ubuntu/Debian
sudo apt-get update
sudo apt-get install docker.io
sudo systemctl start docker
sudo usermod -aG docker $USERFedora/RHEL
sudo dnf install docker
sudo systemctl start docker
sudo usermod -aG docker $USERArch Linux
sudo pacman -S docker
sudo systemctl start docker
sudo usermod -aG docker $USERmacOS
The one-liner install works on all macOS versions 11+.
Optional: Install Docker (for Docker-based skills):
# Download Docker Desktop
# Visit: https://www.docker.com/products/docker-desktop
# Or via Homebrew
brew install --cask dockerWindows (WSL2)
The one-liner install works in WSL2 Ubuntu/Debian.
Optional: Install Docker:
- Download Docker Desktop for Windows with WSL2 backend
- Visit: https://www.docker.com/products/docker-desktop
Claude Code Setup
If you're using Claude Code, set up MCP integration automatically:
skill claude setupThis command:
- Finds your
skillbinary automatically - Creates/updates
.mcp.jsonin your current directory - Configures the MCP server for Claude Code
For global setup (use across all projects):
skill claude setup --globalVerify integration:
skill claude statusOutput:
✓ Claude Code integration configured
Location: /path/to/project/.mcp.json
Server name: skill-engine
Binary: /Users/you/.skill-engine/bin/skillRestart Claude Code and you're ready to use skills through Claude!
See the Claude Code Integration Guide for advanced configuration and tips.
Shell Completion
Enable command-line completion for your shell:
Bash
# Generate completion script
skill completions bash > ~/.local/share/bash-completion/completions/skill
# Or add to ~/.bashrc
echo 'eval "$(skill completions bash)"' >> ~/.bashrc
source ~/.bashrcZsh
# Generate completion script
skill completions zsh > ~/.zfunc/_skill
# Add to ~/.zshrc
echo 'fpath=(~/.zfunc $fpath)' >> ~/.zshrc
echo 'autoload -Uz compinit && compinit' >> ~/.zshrc
source ~/.zshrcFish
# Generate completion script
skill completions fish > ~/.config/fish/completions/skill.fishConfiguration
Default Configuration
Create ~/.config/skill-engine/config.toml:
# Default manifest file location
manifest_path = ".skill-engine.toml"
# Enable execution history
history_enabled = true
history_limit = 1000
# HTTP server settings
[server]
host = "127.0.0.1"
port = 3000
ui_enabled = true
# Docker runtime settings
[docker]
default_memory = "2g"
default_cpus = "4"
network = "bridge"
# WASM runtime settings
[wasm]
max_memory_pages = 256 # 16MB
max_execution_time_ms = 30000Environment Variables
Skill Engine respects these environment variables:
# Configuration file location
export SKILL_ENGINE_CONFIG=~/.config/skill-engine/config.toml
# Manifest file name
export SKILL_ENGINE_MANIFEST=.skill-engine.toml
# Enable debug logging
export RUST_LOG=skill=debug
# Docker host
export DOCKER_HOST=unix:///var/run/docker.sock
# HTTP server port
export SKILL_ENGINE_PORT=3000Add to ~/.bashrc or ~/.zshrc to make permanent.
MCP Server Configuration
For Claude Code and other MCP-compatible agents:
Claude Code
Create or edit ~/.config/claude/mcp.json:
{
"mcpServers": {
"skill-engine": {
"command": "skill",
"args": ["mcp"],
"env": {
"RUST_LOG": "info"
}
}
}
}Cursor
Create or edit ~/.cursor/mcp.json:
{
"mcpServers": {
"skill-engine": {
"command": "skill",
"args": ["mcp"]
}
}
}Verification
Verify your installation is working correctly:
# Check version
skill --version
# List available skills
skill list
# Check system info
skill info
# Start HTTP server
skill serve &
# Test API endpoint
curl http://localhost:3000/api/health
# Stop server
pkill skillTroubleshooting
Command Not Found
Issue: skill: command not found
Solution: Ensure ~/.cargo/bin is in your PATH:
# Add to ~/.bashrc or ~/.zshrc
export PATH="$HOME/.cargo/bin:$PATH"
source ~/.bashrc # or source ~/.zshrcDocker Permission Denied
Issue: permission denied while trying to connect to Docker
Solution: Add user to docker group:
sudo usermod -aG docker $USER
newgrp docker
# Or restart systemRust Version Too Old
Issue: requires rustc 1.75.0 or newer
Solution: Update Rust:
rustup update stable
rustc --versionBuild Errors
Issue: Compilation errors during cargo install
Solution: Update Rust and dependencies:
rustup update stable
cargo clean
cargo build --releasemacOS Gatekeeper Warning
Issue: "skill cannot be opened because it is from an unidentified developer"
Solution: Allow the application:
# Remove quarantine attribute
xattr -d com.apple.quarantine /usr/local/bin/skill
# Or right-click → Open in FinderUninstallation
Remove Binary
# If installed via cargo
cargo uninstall skill-cli
# If installed manually
sudo rm /usr/local/bin/skillRemove Configuration
# Remove config directory
rm -rf ~/.config/skill-engine
# Remove completion scripts
rm ~/.local/share/bash-completion/completions/skill # Bash
rm ~/.zfunc/_skill # Zsh
rm ~/.config/fish/completions/skill.fish # FishRemove Docker Images
# List Skill Engine related images
docker images | grep skill
# Remove specific image
docker rmi <image-name>
# Or remove all unused images
docker image prune -aUpdating
Update via Cargo
cargo install skill-cli --forceUpdate from Source
cd skill
git pull
cargo install --path crates/skill-cli --forceNext Steps
- Quick Start Guide: Run your first skill
- CLI Reference: Learn all commands
- Configuration Guide: Advanced configuration
- Building Skills: Create custom skills
Support
- GitHub Issues: Report bugs
- Discussions: Ask questions
- Documentation: docs.skill-engine.dev