Skip to content

Commands

LLMling Agent provides a consistent command system across both CLI and web interfaces. Commands are prefixed with / and can take arguments and options.

Essential Commands

Session Management

/exit
Exit the current chat session cleanly.

Tool Management

# List available tools
/list-tools

# Enable/disable specific tools
/enable-tool <name>
/disable-tool <name>

# Show detailed tool information
/tool-info <name>

# Register new tools
/register-tool <import_path> [--name custom_name] [--description "Tool description"]
Example: /register-tool json.dumps --name format_json --description "Format JSON data"

# Write and register tools directly in chat
/write-tool

Resource Management

# List available resources
/list-resources

# Show resource content
/show-resource <name> [--param1 value1] [--param2 value2]
Example: /show-resource config.yml
Example: /show-resource template --date today

Agent Configuration

# Show current agent configuration
/show-agent

# Switch to different agent
/switch-agent <name>

# List available agents
/list-agents

# Change model
/set-model <model>
Example: /set-model gpt-4

Additional Commands

Environment

# Change environment file
/set-env <path>

# Open environment configuration
/open-env-file

History and Stats

# Clear chat history
/clear

# Reset session state
/reset

# Search conversation history
/search-history [query] [--hours N] [--limit N]

# Show usage statistics
/show-statistics [--group-by model|agent|hour|day] [--hours N]

Utility Commands

# Copy last assistant message
/copy-clipboard

# Open agent configuration file
/open-agent-file

# Show help
/help [command]

Command Categories

Core Operations

  • /help - Show available commands or help for specific command
  • /exit - Exit chat session
  • /clear - Clear chat history
  • /reset - Reset session state

Tool Management

  • /list-tools - List available tools
  • /enable-tool - Enable a tool
  • /disable-tool - Disable a tool
  • /tool-info - Show tool details
  • /register-tool - Register new tool
  • /write-tool - Create new tool interactively

Resource Management

  • /list-resources - List available resources
  • /show-resource - Show resource content

Agent Management

  • /show-agent - Show current configuration
  • /switch-agent - Switch to different agent
  • /list-agents - List available agents
  • /set-model - Change model

Environment

  • /set-env - Change environment file
  • /open-env-file - Open environment config

History and Analysis

  • /search-history - Search conversation history
  • /show-statistics - Show usage statistics

Utilities

  • /copy-clipboard - Copy last response
  • /open-agent-file - Open config file

Command Line Conventions

  • Commands start with forward slash (/)
  • Arguments can be positional or named: 
    /command arg1 arg2              # Positional
/command --name value           # Named
/command arg --name "value"     # Mixed
  • Arguments with spaces must be quoted
  • Options are prefixed with --

Command Output

Commands provide feedback in different ways: - Direct responses (success/failure messages) - Formatted data (tables, JSON, YAML) - System messages in chat - Status updates

Note: Some commands may be restricted based on agent capabilities and role configuration.