Skip to content

CAI CLI Commands Reference

This comprehensive guide documents all commands available in the CAI Command Line Interface, organized by category for easy navigation.

Command Categories

  1. Agent Management
  2. Model Management
  3. Memory & History
  4. Environment & Configuration
  5. Tools & Integration
  6. System Management
  7. Parallel Execution
  8. Utilities

Agent Management

/agent or /a

Manage and switch between different AI agents.

Syntax: 

/agent [subcommand] [arguments]
/a [subcommand] [arguments]

Subcommands:

list

List all available agents with their descriptions.

/agent list

Output: Table showing agent names, descriptions, and primary use cases.

<agent_name>

Switch to a specific agent.

/agent redteam_agent
/agent bug_bounter_agent
/a blueteam_agent

info or info <agent_name>

Display detailed information about the current or specified agent.

# Current agent info
/agent info

# Specific agent info
/agent info redteam_agent

Examples:

# List all agents
CAI> /agent list

# Switch to red team agent
CAI> /agent redteam_agent

# Switch to bug bounty agent (using alias)
CAI> /a bug_bounter_agent

# Get info about DFIR agent
CAI> /agent info dfir_agent

Available Agents:

Agent Use Case
redteam_agent Offensive security testing
blueteam_agent Defensive security analysis
bug_bounter_agent Bug bounty hunting
one_tool_agent Single-tool execution
dfir_agent Digital forensics
reverse_engineering_agent Binary analysis
network_security_analyzer_agent Network security
wifi_security_agent WiFi security testing
android_sast_agent Android security analysis
selection_agent Agent recommendation

Notes: - Agent changes are immediate - Conversation history is preserved when switching - Each agent has specialized tools and instructions

Model Management

/model or /mod

View or change the current LLM model.

Syntax: 

/model [model_name]
/mod [model_name]

Examples:

# View current model
CAI> /model

# Change to alias1
CAI> /model alias1

# Change to GPT-4o
CAI> /model gpt-4o

# Change to Claude
CAI> /model claude-3-5-sonnet-20241022

/model-show

Display all available models from the LiteLLM repository.

Syntax: 

/model-show

Output: Comprehensive list of models by provider (OpenAI, Anthropic, Ollama, etc.)

Examples:

# Show all available models
CAI> /model-show

# Then select one
CAI> /model gpt-4o

Commonly Used Models:

Model Provider Cost Best For
alias1 Alias Robotics Medium Balanced performance ⭐
gpt-4o OpenAI High Complex reasoning
claude-3-5-sonnet-20241022 Anthropic High Fast & accurate
o1-mini OpenAI Medium Reasoning tasks

Memory & History

/history or /his

Display conversation history.

Syntax: 

/history [number] [agent_name]
/his [number]

Parameters: - number: Number of recent messages to show (default: 10) - agent_name: Filter by specific agent

Examples:

# Show last 10 messages
CAI> /history

# Show last 20 messages
CAI> /history 20

# Show last 5 messages
CAI> /his 5

# Show history for specific agent
CAI> /history 10 redteam_agent

Output: Formatted conversation with timestamps, roles (user/agent), and message content.

/memory or /mem

Manage persistent memory storage across sessions.

Syntax: 

/memory <subcommand> [arguments]
/mem <subcommand> [arguments]

Subcommands:

list

Show all saved memories.

/memory list

save [name]

Save current conversation as a memory.

/memory save "web app pentest findings"
/mem save ctf_techniques

apply <memory_id>

Apply a saved memory to the current session.

/memory apply mem_12345

show <memory_id>

Display the content of a specific memory.

/memory show mem_12345

delete <memory_id>

Remove a memory permanently.

/memory delete mem_12345

merge <id1> <id2> [name]

Combine two memories into one.

/memory merge mem_12345 mem_67890 "combined_findings"

compact

AI-powered memory summarization.

/memory compact

status

Show memory system status and statistics.

/memory status

Examples:

# Save current session insights
CAI> /memory save "SQLi vulnerabilities found"

# List all memories
CAI> /memory list

# Apply previous knowledge
CAI> /memory apply mem_12345

# Check memory status
CAI> /mem status

Notes: - Memories persist across sessions - Stored in .cai/memory/ directory - Useful for long-term research projects

/compact or /cmp

Compact the current conversation to reduce context size.

Syntax: 

/compact [model_name]
/cmp

Parameters: - model_name: Optional model to use for compaction

Examples:

# Compact with current model
CAI> /compact

# Compact with specific model
CAI> /compact alias1

Use Cases: - Approaching token limits - Long conversations that need summarization - Maintaining conversation flow with reduced tokens

/flush or /clear

Clear conversation history.

Syntax: 

/flush [agent_name|all]
/clear

Parameters: - agent_name: Flush specific agent history - all: Flush all agent histories

Examples:

# Flush current agent
CAI> /flush

# Flush specific agent
CAI> /flush redteam_agent

# Flush all agents
CAI> /flush all

Warning: This action is irreversible. Consider using /save first.

/load or /l

Load conversation history from a file.

Syntax: 

/load <filename>
/l <filename>

Supported Formats: - JSON (.json) - JSONL (.jsonl) - Markdown (.md)

Examples:

# Load JSON session
CAI> /load pentest_session.json

# Load JSONL data
CAI> /load conversation.jsonl

# Using alias
CAI> /l ~/sessions/previous_work.json

Notes: - Restores conversation context - Compatible with /save output - Can load partial histories

/merge or /mrg

Merge agent message histories (shortcut for /parallel merge).

Syntax: 

/merge [agent1] [agent2]
/mrg

Examples:

# Merge all parallel agents
CAI> /merge

# Merge specific agents
CAI> /merge redteam_agent blueteam_agent

Use Cases: - Combining parallel execution results - Integrating different agent perspectives

Environment & Configuration

/config or /cfg

Display and configure environment variables.

Syntax: 

/config [VARIABLE=value]
/config set <number> <value>
/cfg

Examples:

# View all configuration
CAI> /config

# Set by variable name
CAI> /config CAI_PRICE_LIMIT=10.0
CAI> /config CAI_MAX_TURNS=50

# Set by number (from /config output)
CAI> /config set 18 "5.0"

Common Configuration Variables:

Variable Description Default
CAI_MODEL Default model alias1
CAI_AGENT_TYPE Default agent redteam_agent
CAI_DEBUG Debug level (0-2) 1
CAI_PRICE_LIMIT Cost limit (USD) 1.0
CAI_MAX_TURNS Max conversation turns inf
CAI_MAX_INTERACTIONS Max tool calls inf
CAI_TRACING Enable tracing true
CAI_GUARDRAILS Security guardrails false

Notes: - Changes take effect immediately - Use /config without arguments to see all options - Numbers in first column can be used with set subcommand

/env or /e

Display current environment variables.

Syntax: 

/env [pattern]
/e

Parameters: - pattern: Optional filter pattern (e.g., "CAI", "CTF")

Examples:

# Show all environment variables
CAI> /env

# Filter CAI-specific variables
CAI> /env CAI

# Filter CTF variables
CAI> /env CTF

/workspace or /ws

Manage workspace directories.

Syntax: 

/workspace <subcommand> [path]
/ws <subcommand>

Subcommands:

show or pwd

Display current workspace directory.

/workspace show
/ws pwd

set <path>

Change workspace directory.

/workspace set /path/to/project
/ws set ~/ctf_challenges

list or ls

List workspace contents.

/workspace list
/ws ls

Examples:

# Show current workspace
CAI> /workspace show

# Change workspace
CAI> /workspace set /home/user/pentests

# List files
CAI> /ws ls

Notes: - Affects where shell commands execute - Useful for CTF challenges and projects - Works with Docker containers

/virtualization or /virt

Manage Docker-based virtualization environments.

Syntax: 

/virtualization <subcommand> [arguments]
/virt <subcommand>

Subcommands:

list

List available containers.

/virtualization list

set <container_id>

Set active container for command execution.

/virtualization set abc123def456
/virt set mycontainer

clear

Return to host environment.

/virtualization clear

info

Show current virtualization status.

/virtualization info

Examples:

# List containers
CAI> /virtualization list

# Execute commands in container
CAI> /virt set ubuntu_ctf

# Return to host
CAI> /virt clear

Notes: - Automatically set when CTF challenges start - Commands execute inside specified container - Uses CAI_ACTIVE_CONTAINER environment variable

Tools & Integration

/mcp or /m

Manage Model Context Protocol (MCP) servers and their tools.

Syntax: 

/mcp <subcommand> [arguments]
/m <subcommand>

Subcommands:

load <url> <name>

Load an SSE MCP server.

/mcp load http://localhost:9876/sse burp

load stdio <command> <name>

Load a STDIO MCP server.

/mcp load stdio "npx -y @modelcontextprotocol/server-brave-search" brave

list

List active MCP connections.

/mcp list

add <agent_name> <server_name>

Add MCP tools to an agent.

/mcp add redteam_agent burp

remove <server_name>

Remove an MCP server connection.

/mcp remove burp

tools <server_name>

List tools from an MCP server.

/mcp tools burp

status

Check MCP server connection status.

/mcp status

associations

Show agent-MCP associations.

/mcp associations

Examples:

# Load Burp Suite MCP server
CAI> /mcp load http://localhost:9876/sse burp

# List MCP tools
CAI> /mcp tools burp

# Add to current agent
CAI> /mcp add redteam_agent burp

# Check status
CAI> /mcp status

Common MCP Servers: - Burp Suite: Web application testing tools - Brave Search: Web search capabilities - Filesystem: File operations - Git: Repository management - Postgres: Database operations

Notes: - Extends agent capabilities dynamically - Supports both SSE and STDIO protocols - See MCP Documentation for details

/shell or /s or /$

Execute shell commands directly from the CLI.

Syntax: 

/shell <command>
/s <command>
$ <command>

Examples:

# Using /shell
CAI> /shell nmap -sV 192.168.1.1

# Using /s alias
CAI> /s whoami

# Using $ shortcut
CAI> $ ls -la

# Complex commands
CAI> $ nmap -sV -p- 192.168.1.0/24 -oN scan_results.txt

Notes: - Commands execute in current workspace - Respects CAI_ACTIVE_CONTAINER if set - Output displayed in real-time - Ctrl+C to interrupt running commands

System Management

/kill or /k

Terminate active processes or stuck sessions.

Syntax: 

/kill
/k

Examples:

# Kill current process
CAI> /kill

# Alternative: Ctrl+C

Use Cases: - Stopping stuck tool executions - Canceling long-running operations - Interrupting agent loops

/exit or /quit or /q

Exit the CAI CLI.

Syntax: 

/exit
/quit
/q

Examples:

# Exit CAI
CAI> /exit

# Alternative: Ctrl+D

Notes: - Performs clean shutdown - Saves session logs - Stops background processes

/quickstart

Display setup information and quick start guide.

Syntax: 

/quickstart

Examples:

# Show quickstart guide
CAI> /quickstart

Notes: - Auto-displays on first launch - Useful for new users - Shows essential commands and setup

Parallel Execution

/parallel or /par or /p

Manage parallel agent configurations and execution.

Syntax: 

/parallel <subcommand> [arguments]
/par <subcommand>
/p <subcommand>

Subcommands:

add <agent_name> [model]

Add an agent to parallel configuration.

/parallel add redteam_agent alias1
/par add bug_bounter_agent gpt-4o

remove <agent_id>

Remove an agent from parallel configuration.

/parallel remove P1

list

List all parallel agents.

/parallel list

clear

Clear all parallel configurations.

/parallel clear

run <prompt>

Execute a prompt across all parallel agents.

/parallel run "scan 192.168.1.1 for vulnerabilities"

merge

Merge all parallel agent histories.

/parallel merge

Examples:

# Configure parallel agents
CAI> /parallel add redteam_agent alias1
CAI> /parallel add blueteam_agent alias1
CAI> /parallel add bug_bounter_agent gpt-4o

# List configuration
CAI> /parallel list

# Execute on all agents
CAI> /parallel run "analyze target.com"

# Merge results
CAI> /parallel merge

# Clear configuration
CAI> /parallel clear

YAML Configuration:

Create agents.yaml:

agents:
  - name: red1
    agent_type: redteam_agent
    model: alias1
  - name: bug1
    agent_type: bug_bounter_agent
    model: alias1

Launch with YAML:

cai --yaml agents.yaml --prompt "scan target.com"

Notes: - Each agent runs independently - Results can be merged - Different models per agent supported - See Advanced Usage for more details

/run or /r

Execute queued prompts (works with parallel mode).

Syntax: 

/run <prompt>
/r <prompt>

Examples:

# Queue and run prompt
CAI> /run "analyze this binary"

# Alternative
CAI> /r "test for XSS"

Notes: - Executes immediately if agents are ready - Queues if agents are busy - Works with both single and parallel modes

Queue File Format (prompts.txt):

# Comments start with #
/agent redteam_agent
Scan 192.168.1.0/24 for open ports
Test https://target.com for vulnerabilities
$ nmap -sV 192.168.1.1
Generate security report

Notes: - Prompts execute sequentially - Supports commands and regular prompts - Can load from files for automation

Utilities

/help or /h or /?

Display help information and command documentation.

Syntax: 

/help [command]
/h [command]
/? [command]

Examples:

# General help
CAI> /help

# Help for specific command
CAI> /help agent
CAI> /h parallel
CAI> /? mcp

Topics: - agent: Agent management - parallel: Parallel execution - memory: Memory management - config: Configuration - mcp: MCP integration - commands: List all commands

/graph or /g

Visualize agent interaction graphs.

Syntax: 

/graph [agent_name]
/g

Examples:

# Show graph for current conversation
CAI> /graph

# Show graph for specific agent
CAI> /graph redteam_agent

Output: - Directed graph of conversations - User and agent interactions - Tool calls highlighted - Conversation flow visualization

/context or /ctx 🚀 CAI PRO Exclusive

⚡ CAI PRO Exclusive Feature
The /context command is available exclusively in CAI PRO. To access this feature and unlock advanced monitoring capabilities, visit Alias Robotics for more information.

View context usage and token statistics for the current conversation.

Syntax: 

/context [agent_name]
/ctx

Examples:

# Show context for current agent
CAI> /context

# Show context for specific agent
CAI> /ctx redteam_agent

Output Includes: - Total context usage (used/max tokens) with percentage - Visual grid representation with CAI logo - Breakdown by category: - System prompt tokens - Tool definitions tokens - Memory/RAG tokens - User prompts tokens - Assistant responses tokens - Tool calls tokens - Tool results tokens - Free space available - Color-coded visualization

Notes: - Helps monitor token limits - Useful for long conversations - Different models have different context windows

/cost

Display API usage costs and token statistics.

Syntax: 

/cost [agent_name]

Examples:

# Show costs for current session
CAI> /cost

# Show costs for specific agent
CAI> /cost redteam_agent

# Show all agents' costs
CAI> /cost all

Output Includes: - Total cost (USD) - Input tokens used - Output tokens used - Cost per interaction - Model pricing rates - Agent breakdown

/save

Save current conversation to a file.

Syntax: 

/save <filename>

Supported Formats: - JSON (.json) - Markdown (.md)

Examples:

# Save as JSON
CAI> /save pentest_session.json

# Save as Markdown
CAI> /save findings_report.md

# Full path
CAI> /save ~/sessions/project_alpha.json

Notes: - Saves all conversation history - Includes agent names and timestamps - Cost information preserved - Can be loaded with /load

/temperature or /temp

Adjust the model's temperature parameter.

Syntax: 

/temperature <value>
/temp <value>

Parameters: - value: Temperature (0.0 - 2.0) - Lower = more deterministic - Higher = more creative

Examples:

# Set to more deterministic
CAI> /temperature 0.2

# Set to more creative
CAI> /temp 1.5

# View current temperature
CAI> /temperature

/api

Manage API keys and authentication.

Syntax: 

/api <subcommand> [arguments]

Subcommands: - show: Display configured API keys (masked)

Examples:

# Show API keys
CAI> /api show

---

## Special Features

### Command Chaining

Chain multiple commands using semicolons (`;`).

**Syntax**:
```bash
command1 ; command2 ; command3

Examples:

# Chain commands at launch
cai --prompt "/agent redteam_agent ; scan 192.168.1.1 ; /save results.json"

# Chain in CLI
CAI> /agent bug_bounter_agent ; test https://target.com ; /cost

Use Cases: - Automation workflows - Batch operations - Quick sequences

Auto-loading Queue from File

Load and execute prompts automatically on startup.

Environment Variable: 

export CAI_QUEUE_FILE="/path/to/prompts.txt"

Launch: 

CAI_QUEUE_FILE=~/my_prompts.txt cai

Notes: - Prompts execute automatically - Returns to interactive mode when done - Perfect for automation

Quick Reference

Most Used Commands

Command Description Example
/agent <name> Switch agent /agent redteam_agent
/model <name> Change model /model alias1
/config View config /config
/help Get help /help agent
/save <file> Save session /save session.json
/load <file> Load session /load session.json
/cost Show costs /cost
/history View history /history 20
$ <cmd> Shell command $ nmap -sV target
/exit Exit CAI /exit

Command Aliases

Full Command Aliases
/agent /a
/model /mod
/config /cfg
/help /h, /?
/history /his
/memory /mem
/workspace /ws
/virtualization /virt
/parallel /par, /p
/shell /s, /$
/context /ctx
/compact /cmp
/temperature /temp
/load /l
/merge /mrg
/run /r
/kill /k
/exit /quit, /q

Next Steps

Last updated: November 2025 | CAI CLI v0.6+