Home › Usage & Help

Usage & Help

Complete CLI reference for ScorpioX Code — command-line flags, interactive commands, slash commands, configuration, and quick-start examples.

19
Flags
16
CLI Commands
13
Config Keys
∞
Slash Commands

Quick Start

1

Install ScorpioX Code

curl -fsSL "https://get.scorpiox.net?platform=linux" | bash
2

Start an interactive session

sx
3

Or start with a prompt

sx "refactor the auth module to use JWT tokens"
4

Initialize project configuration

sx --init

CLI Flags

Flag Description
sx Launch ScorpioX Code interactive session
sx "prompt" Start session with an initial prompt
sx -p provider Select AI provider (claude_code, openai, gemini, etc.)
sx -m model Select model name for the provider
sx -c Continue previous session
sx -r Resume last session and replay context
sx --thinking Enable extended thinking / chain-of-thought
sx --no-thinking Disable extended thinking
sx --verbose Enable verbose output (debug logging)
sx --version Print version and build information
sx --help Show help message and exit
sx --init Initialize project with CLAUDE.md and .scorpiox/
sx --agents List available agents
sx --mcp List MCP server configurations
sx --tools List all available tools
sx --plan Enter plan mode for complex tasks
sx --yolo Auto-approve all tool calls (no confirmation prompts)
sx --agentless Run in agentless mode (no sub-agents)
sx --compact Use compact TUI layout

Interactive Commands

Command Description
/help Show available commands and key bindings
/clear Clear the conversation and start fresh
/compact Toggle compact output mode
/plan Enter plan mode — create structured plans before execution
/tasks Manage persistent tasks across sessions
/rewind Undo the last assistant turn
/voice Toggle voice input mode
/screenshot Capture and analyze a screenshot
/search query Search the web for information
/docs URL Fetch and read documentation from a URL
/hook name Trigger a registered hook script
/init Initialize project configuration
/mcp Manage MCP server connections
/agent name Switch to or spawn a sub-agent
/logs Print session logs
/exit End the session

Slash Commands

Slash commands are custom executable scripts placed in .scorpiox/commands/ (project-level) or ~/.scorpiox/commands/ (global). Any executable file becomes a command — bash, python, node, or compiled binaries. Arguments are passed via the $ARGUMENTS environment variable.
/command-name args Execute a custom slash command defined in .scorpiox/commands/
Location: .scorpiox/commands/ Project-level custom commands directory
Location: ~/.scorpiox/commands/ Global custom commands directory
Format: executable script Any executable file — bash, python, node, etc.
$ARGUMENTS Arguments passed to the command script via environment variable

Usage Examples

Provider & Model Selection 
# Use Claude with thinking enabled
sx -p claude_code --thinking "analyze this codebase"

# Use OpenAI GPT-4
sx -p openai -m gpt-4o "write unit tests"

# Use Gemini
sx -p gemini -m gemini-2.5-pro "review PR #42"
Session Management 
# Continue the last session
sx -c

# Resume with full context replay
sx -r

# Auto-approve all tool calls (YOLO mode)
sx --yolo "fix all lint errors in src/"
Configuration via scorpiox-env.txt 
# ~/.scorpiox/scorpiox-env.txt
SX_PROVIDER=claude_code
SX_THINKING=true
USAGE_TRACKING=0
CONTEXT_WINDOW_SIZE=200000
PERMISSIONS_WRITABLE_PATHS=.
TOOL_TIMEOUT=120
Custom Slash Command Example 
# Create a custom command: .scorpiox/commands/deploy
#!/bin/bash
# Deploy to staging environment
echo "Deploying $ARGUMENTS to staging..."
git push origin main
ssh staging "cd /app && git pull && restart"

# Use it in session:
/deploy v2.1.0

Configuration

Configuration is stored in scorpiox-env.txt as key=value pairs. Searched in three locations: <exe_dir>/scorpiox-env.txt, ~/.scorpiox/scorpiox-env.txt, and ./.scorpiox/scorpiox-env.txt (project-level overrides).
Variable Default Description
SX_PROVIDER claude_code Default AI provider
SX_MODEL (provider default) Default model for the provider
SX_THINKING 0 Enable thinking by default (0 or 1)
SX_YOLO 0 Auto-approve all tool calls (0 or 1)
AGENTLESS_MODE 0 Run without sub-agents (0 or 1)
USAGE_TRACKING 0 Enable usage/telemetry tracking
USAGE_API_URL — API endpoint for usage tracking data
EMIT_SESSION_TRACKING 0 Enable session event emission
EMIT_SESSION_API_URL — API endpoint for session event data
CONTEXT_WINDOW_SIZE 200000 Maximum context window tokens
TOOL_TIMEOUT 120 Tool execution timeout in seconds
TMUX_MODE — Container backend mode (podman, docker, etc.)
PERMISSIONS_WRITABLE_PATHS . Paths the agent can write to

Environment Variables

These environment variables can be set in your shell or scorpiox-env.txt to control behavior. Project-level settings in .scorpiox/scorpiox-env.txt override global settings.
Essential Environment Variables 
# Provider API keys
ANTHROPIC_API_KEY=sk-ant-...
OPENAI_API_KEY=sk-...
GEMINI_API_KEY=AI...

# ScorpioX Code settings
SX_PROVIDER=claude_code
SX_MODEL=opus
SX_THINKING=true

# Permissions and safety
PERMISSIONS_WRITABLE_PATHS=./src,./tests
TOOL_TIMEOUT=120

# Container / sandbox mode
TMUX_MODE=podman
TMUX_PODMAN_IMAGE=scorpiox-dev:latest