Agent Commands

Complete reference for AI agent management and execution commands

Agent Commands

NikCLI provides powerful AI agent management capabilities for autonomous task execution. These commands allow you to list, configure, and execute tasks with specialized AI agents.

Core Agent Commands

/agents

List all available AI agents with their capabilities and status.

Syntax:

/agents

Output includes:

• Agent name and type
• Specialization areas
• Current status (available/busy)
• Performance metrics
• Last used timestamp

Example:

/agents

Sample Output:

🤖 Available AI Agents:

┌─────────────────┬──────────────┬────────────┬─────────────┐
│ Agent           │ Type         │ Status     │ Specializes │
├─────────────────┼──────────────┼────────────┼─────────────┤
│ backend-expert  │ Specialized  │ Available  │ APIs, DBs   │
│ frontend-expert │ Specialized  │ Available  │ React, UI   │
│ devops-expert   │ Specialized  │ Busy       │ CI/CD, K8s  │
│ universal-agent │ Universal    │ Available  │ General     │
│ autonomous-coder│ Autonomous   │ Available  │ Full-stack  │
└─────────────────┴──────────────┴────────────┴─────────────┘

/agent [name] [task]

Execute a specific task using a named agent.

Syntax:

/agent <agent-name> "<task-description>"

Parameters:

• agent-name - Name of the agent to use
• task-description - Detailed description of the task

Available Agents:

• backend-expert - Backend development, APIs, databases
• frontend-expert - Frontend development, React, UI/UX
• devops-expert - DevOps, CI/CD, containerization
• code-review - Code review and quality analysis
• system-admin - System administration and configuration
• universal-agent - General-purpose agent
• autonomous-coder - Fully autonomous coding

Examples:

# Backend development
/agent backend-expert "Create a REST API for user authentication with JWT tokens"

# Frontend development
/agent frontend-expert "Build a responsive dashboard with charts using React and Tailwind"

# DevOps tasks
/agent devops-expert "Set up CI/CD pipeline with GitHub Actions and Docker"

# Code review
/agent code-review "Review the authentication module for security issues"

# System administration
/agent system-admin "Configure nginx reverse proxy with SSL certificates"

# General tasks
/agent universal-agent "Analyze this codebase and suggest improvements"

/auto [task]

Execute a task with full autonomous mode - the system selects the best agent and approach.

Syntax:

/auto "<task-description>"

Features:

• Automatic agent selection
• Multi-step task breakdown
• Progress monitoring
• Error recovery
• Resource optimization

Examples:

# Full application development
/auto "Build a todo app with React frontend and Node.js backend"

# Complex refactoring
/auto "Migrate this Express app to use TypeScript with proper types"

# Infrastructure setup
/auto "Deploy this application to AWS with proper monitoring and scaling"

# Bug investigation
/auto "Find and fix the memory leak in the user service"

# Performance optimization
/auto "Optimize this React app for better performance and SEO"

/parallel [tasks]

Execute multiple tasks in parallel using different agents.

Syntax:

/parallel "<task1>" "<task2>" "<task3>"

Features:

• Concurrent execution
• Resource management
• Progress tracking
• Dependency handling
• Result aggregation

Examples:

# Development workflow
/parallel "Run tests" "Build production bundle" "Update documentation"

# Code quality
/parallel "Run linter" "Check security vulnerabilities" "Analyze performance"

# Deployment tasks
/parallel "Build Docker image" "Update Kubernetes manifests" "Run integration tests"

# Multi-service development
/parallel "Update user service API" "Update frontend components" "Update database schema"

Agent Factory Commands

/factory

Show the agent factory status and available agent templates.

Syntax:

/factory

Output includes:

• Available agent templates
• Custom agent definitions
• Factory configuration
• Resource usage
• Creation history

/create-agent [specification]

Create a custom agent with specific capabilities.

Syntax:

/create-agent <agent-spec>

Agent Specifications:

• frontend - Frontend-focused agent
• backend - Backend-focused agent
• fullstack - Full-stack development agent
• mobile - Mobile development agent
• data - Data analysis and ML agent
• security - Security-focused agent
• custom - Interactive custom agent creation

Examples:

# Create frontend specialist
/create-agent frontend

# Create backend specialist
/create-agent backend

# Create custom agent interactively
/create-agent custom

Interactive Creation Process:

1. Agent name and description
2. Specialization areas
3. Tool access permissions
4. Resource limits
5. Behavioral parameters

/launch-agent [name]

Launch a previously created custom agent.

Syntax:

/launch-agent <agent-name>

Examples:

# Launch custom agent
/launch-agent my-react-expert

# Launch with specific task
/launch-agent my-backend-agent "optimize database queries"

Advanced Agent Features

Agent Streaming

Real-time output streaming from agent execution:

# Enable streaming for agent commands
/stream enable

# Execute with streaming
/agent backend-expert "create API endpoints"
# Output streams in real-time as agent works

Agent Context

Agents maintain context across interactions:

# Set context for subsequent agent calls
/context set "working on e-commerce project with React and Node.js"

# Agent calls will use this context
/agent frontend-expert "add shopping cart functionality"

Agent Memory

Agents can remember information across sessions:

# Agent learns from interactions
/agent backend-expert "create user model with validation"

# Later, agent remembers previous work
/agent backend-expert "add authentication to the user model"

Agent Configuration

Model Selection

Configure which AI model agents use:

# Set model for all agents
/model claude-3-5-sonnet

# Set model for specific agent type
/router set backend-expert gpt-4

# View current model configuration
/router status

Agent Permissions

Configure what agents can access:

# Security settings
/security agent-permissions

# Enable/disable specific capabilities
/dev-mode enable  # Allows broader access
/safe-mode        # Restricts agent capabilities

Resource Limits

Set limits on agent resource usage:

# Configure in system settings
/system agent-limits

# View current limits
/stats agents

Agent Workflows

Development Workflow

# 1. Plan the work
/plan "implement user authentication system"

# 2. Execute with appropriate agent
/agent backend-expert "create JWT authentication middleware"

# 3. Test and validate
/agent code-review "review authentication implementation"

# 4. Deploy
/agent devops-expert "deploy authentication service"

Code Review Workflow

# 1. Analyze current code
/agent code-review "analyze codebase for issues"

# 2. Get specific recommendations
/agent code-review "suggest performance improvements"

# 3. Implement fixes
/auto "implement the suggested performance improvements"

# 4. Validate changes
/agent code-review "verify performance improvements"

Multi-Agent Collaboration

# 1. Parallel development
/parallel "backend API development" "frontend component creation" "database schema design"

# 2. Integration
/agent universal-agent "integrate frontend with backend API"

# 3. Testing
/parallel "unit tests" "integration tests" "e2e tests"

# 4. Deployment
/agent devops-expert "deploy integrated application"

Agent Troubleshooting

Common Issues

Agent not responding:

# Check agent status
/agents

# Restart agent system
/factory reset

# Check system resources
/debug agents

Poor agent performance:

# Check model configuration
/model

# Verify API keys
/config

# Check resource limits
/system limits

Agent errors:

# View detailed logs
/debug agent-logs

# Check permissions
/security status

# Verify tool access
/tools status

Debug Commands

# Show agent debug information
/debug agents

# Show agent execution logs
/debug agent-logs [agent-name]

# Show agent performance metrics
/stats agents

# Test agent connectivity
/diagnostic agents

Best Practices

Task Description

Write clear, specific task descriptions:

# ✅ Good: Specific and actionable
/agent backend-expert "Create a REST API endpoint for user registration that validates email format, checks for duplicate emails, hashes passwords with bcrypt, and returns JWT tokens"

# ❌ Bad: Vague and unclear
/agent backend-expert "make user stuff work"

Agent Selection

Choose the right agent for the task:

# ✅ Use specialized agents for their expertise
/agent frontend-expert "optimize React component rendering"
/agent devops-expert "configure Kubernetes ingress"
/agent security-expert "audit authentication flow"

# ✅ Use universal agent for general tasks
/agent universal-agent "analyze project structure"

# ✅ Use autonomous mode for complex multi-step tasks
/auto "build complete user management system"

Resource Management

Monitor and manage agent resource usage:

# Check current usage
/stats agents

# Set appropriate limits
/system agent-limits

# Use parallel execution efficiently
/parallel "quick task 1" "quick task 2"  # Good
# Avoid: /parallel "long task 1" "long task 2" "long task 3"  # May overwhelm system

Error Handling

Handle agent errors gracefully:

# If agent fails, try different approach
/agent backend-expert "create API"  # If this fails
/auto "create API"                  # Try autonomous mode
/agent universal-agent "create API" # Or use universal agent

Integration Examples

CI/CD Integration

# In CI pipeline
/agent code-review "analyze pull request changes"
/parallel "run tests" "check security" "validate performance"
/agent devops-expert "deploy if all checks pass"

Development Environment

# Daily development workflow
/agent universal-agent "analyze today's tasks from project board"
/auto "implement highest priority feature"
/agent code-review "review completed work"
/agent devops-expert "deploy to staging environment"

Code Maintenance

# Regular maintenance
/agent code-review "identify technical debt"
/agent backend-expert "optimize database queries"
/agent frontend-expert "update dependencies and fix deprecations"
/parallel "update documentation" "run security audit" "performance testing"