Agentic CLI

An advanced AI-powered CLI tool combining the best of Claude Code with cutting-edge features: multi-agent orchestration, security scanning, web search integration, automated testing, codebase indexing, and intelligent workflow automation.

Features

Core Capabilities

• Multi-Agent Support: Run parallel tasks with intelligent orchestration
• AI-Powered Code Editing: Multi-file edits with full context awareness
• Semantic Search: Natural language code search across repositories
• Git Integration: Auto-commit, push, and PR creation with AI
• Plugin Architecture: Extensible with custom commands and hooks
• Multiple UI Modes: CLI, Terminal UI (TUI), or quiet mode

Advanced Features

• Security Scanning: Built-in vulnerability detection with auto-fixing
• Web Search Integration: Live web search for grounded AI answers
• Session Management: Persistent context with full audit logging
• Test Generation: Automated test creation with coverage analysis
• Codebase Indexing: Symbol tracking and dependency graphs
• Workflow Automation: Multi-step orchestration with conditions

Animated User Interface

• Typewriter Effects: Commands and responses typed out character by character
• Progress Animations: Smooth progress bars and loading indicators
• Agent Status: Real-time animated status indicators for all agents
• Banner Animations: Engaging startup sequence with slide-in effects
• Transition Effects: Slide, fade, and pulse animations for enhanced UX
• Customizable Speed: Adjust animation speed from 1 (slowest) to 10 (fastest)

Specialized Review Agents

• comment-analyzer: Documentation quality and accuracy
• pr-test-analyzer: Test coverage and behavioral testing
• silent-failure-hunter: Error handling and failure detection
• type-design-analyzer: Type system quality (TypeScript/Python)
• code-reviewer: Project guidelines and bug detection
• code-simplifier: Code clarity and refactoring

Installation

Quick Install (Recommended)

# Install from npm
npm install -g @judeotine/agentic-cli

# Initialize configuration
agentic init

Development Install

# Clone the repository
git clone https://github.com/judeotine/Agentic-CLI.git
cd Agentic-CLI

# Install dependencies
npm install

# Build the project
npm run build

# Link for global use
npm link

# Initialize configuration
agentic init

Quick Start

1. Configure API Keys

Create .env file:

# At least one AI provider key required
OPENAI_API_KEY=sk-...
# OR
ANTHROPIC_API_KEY=sk-ant-...

# Optional: Web search (for enhanced features)
BRAVE_API_KEY=...

# Optional: GitHub integration
GITHUB_TOKEN=ghp_...

2. Initialize

agentic init

3. Animation Demo

Experience the full animation system:

# Full animation showcase
agentic demo

# Custom speed (1=slowest, 10=fastest)
agentic demo --speed 3

# Skip specific animations
agentic demo --no-banner --no-progress

The demo showcases:

• Animated banner entrance
• Typewriter command effects
• Progress bar animations
• Agent status indicators
• Slide and fade transitions
• Pulse effects for notifications

4. Start Developing

# Guided feature development (Claude Code workflow)
agentic feature-dev "Add user authentication"

# AI-powered code editing
agentic code-edit src/index.ts --prompt "Add error handling"

# Security scan
agentic security scan --auto-fix

# Generate tests
agentic test generate src/**/*.ts

# Comprehensive PR review
agentic review-pr

# Commit and create PR
agentic commit-push-pr

Command Overview

Development Workflow

agentic feature-dev        # Guided feature development (7 phases)
agentic code-edit          # AI code editing
agentic code-edit:multi    # Multi-file parallel editing

Code Quality

agentic review-pr          # Comprehensive PR review
agentic review-tests       # Test coverage review
agentic review-security    # Security and error handling review
agentic security scan      # Vulnerability scanning
agentic test generate      # Auto-generate tests

Search & Navigation

agentic search             # Text/regex search
agentic search:semantic    # AI semantic search
agentic index build        # Build code index
agentic index find         # Find symbols
agentic index refs         # Find references
agentic web search         # Web search
agentic web ask            # Grounded AI answers

Git & Automation

agentic commit-push-pr     # One-command ship
agentic smart-commit       # AI commit message
agentic git:commit --auto  # Conventional commits
agentic git:diff --analyze # AI diff analysis
agentic clean-gone         # Clean dead branches

Agents & Orchestration

agentic agent:run          # Run single task
agentic agent:parallel     # Run parallel tasks
agentic agent:create       # Create custom agent
agentic agent:status       # Monitor agents

Sessions & Context

agentic session create     # New session
agentic session list       # List sessions
agentic session export     # Export session
agentic session audit      # Audit log

Workflows

agentic workflow create    # Create workflow
agentic workflow run       # Execute workflow
agentic workflow list      # List workflows

Example Workflows

Complete Feature Development

# Start session
agentic session create "oauth-feature"

# Guided development with exploration and architecture
agentic feature-dev "Add OAuth2 authentication with Google and GitHub"

# Follow 7-phase workflow:
# 1. Discovery - Clarify requirements
# 2. Exploration - 3 code-explorer agents analyze codebase
# 3. Questions - Answer all ambiguities
# 4. Architecture - 3 code-architect agents propose approaches
# 5. Implementation - Build chosen architecture
# 6. Review - 3 code-reviewer agents validate
# 7. Summary - Document what was built

# Generate comprehensive tests
agentic test generate src/auth/**/*.ts --coverage 90

# Security validation
agentic security scan --auto-fix

# Final review
agentic review-pr

# Ship it
agentic commit-push-pr

Security-First Development

# Scan before starting
agentic security scan --severity high

# Edit with security hooks active (warns in real-time)
agentic code-edit src/api/**/*.ts --prompt "Add rate limiting"

# Security-focused review
agentic review-security

# Generate security report
agentic security report --format html --output security-report.html

# Commit only after validation
agentic smart-commit

AI-Assisted Code Review

# Index codebase for better context
agentic index build

# Comprehensive multi-agent review
agentic review-pr

# Specific reviews:
agentic review-tests              # Test coverage
agentic review-security           # Error handling
agentic agent:run code-simplifier "Simplify complex functions"

# Web search for best practices
agentic web ask "Best practices for async error handling in Node.js"

# Apply improvements
agentic code-edit src/**/*.ts --prompt "Apply review suggestions"

Project Structure

Agentic-CLI/
├── src/
│   ├── commands/          # 11 command groups
│   │   ├── code-edit.ts
│   │   ├── repo-search.ts
│   │   ├── agent-manager.ts
│   │   ├── git.ts
│   │   ├── security.ts
│   │   ├── test.ts
│   │   ├── session.ts
│   │   ├── web.ts
│   │   ├── index-cmd.ts
│   │   └── workflow.ts
│   ├── core/              # 8 core services
│   │   ├── config.ts
│   │   ├── ai-provider.ts
│   │   ├── agent.ts
│   │   ├── plugin-manager.ts
│   │   ├── git-service.ts
│   │   ├── security-scanner.ts
│   │   ├── web-search.ts
│   │   ├── session-manager.ts
│   │   ├── test-generator.ts
│   │   ├── codebase-indexer.ts
│   │   ├── workflow-engine.ts
│   │   └── prompt-loader.ts
│   ├── ui/
│   │   ├── logger.ts
│   │   └── tui.ts
│   ├── types/
│   │   ├── config.ts
│   │   ├── plugin.ts
│   │   ├── agent.ts
│   │   └── command.ts
│   └── utils/
│       ├── file-ops.ts
│       ├── async.ts
│       └── validation.ts
│
├── plugins/               # Plugin ecosystem
│   ├── feature-development/     # Claude Code integration
│   │   ├── agents/
│   │   │   ├── code-architect.md
│   │   │   ├── code-explorer.md
│   │   │   └── code-reviewer.md
│   │   └── commands/
│   │       └── feature-dev.md
│   ├── pr-review-toolkit/       # 6 specialized agents
│   │   └── agents/
│   │       ├── comment-analyzer.md
│   │       ├── pr-test-analyzer.md
│   │       ├── silent-failure-hunter.md
│   │       ├── type-design-analyzer.md
│   │       └── code-simplifier.md
│   ├── commit-automation/       # Git workflow automation
│   ├── security-hooks/          # Real-time security validation
│   └── example-plugin/
│
├── tests/
├── config/
├── examples/
└── workflows/

Configuration

Config stored at ~/.agentic/config.yaml:

version: '1.0.0'

aiModels:
  gpt-4:
    provider: openai
    model: gpt-4-turbo-preview
    apiKey: ${OPENAI_API_KEY}
    temperature: 0.7
  
  claude-3:
    provider: anthropic
    model: claude-3-sonnet-20240229
    apiKey: ${ANTHROPIC_API_KEY}
    temperature: 0.7

defaultModel: claude-3

plugins:
  enabled: true
  directory: ./plugins
  
  feature-development:
    explorers: 3
    architects: 3
    reviewers: 3
    auto_test: true
    auto_security: true
  
  security-hooks:
    blockOnCritical: false
    warnOnHigh: true

agents:
  maxConcurrent: 5
  timeout: 300000
  retryAttempts: 3

permissions:
  fileSystem:
    read: true
    write: true
    delete: false
  git:
    read: true
    commit: true
    push: true
    createPR: true

Architecture

Unified AI Integration

Combines Claude Code's agent system with Agentic CLI's provider abstraction:

// Supports multiple AI providers
AIProvider
  ├── OpenAIProvider (GPT-4, GPT-3.5)
  ├── AnthropicProvider (Claude 3 Sonnet/Opus)
  ├── AzureProvider
  └── LocalProvider (Ollama, LM Studio)

// With specialized agents
Agents
  ├── code-architect
  ├── code-explorer
  ├── code-reviewer
  ├── silent-failure-hunter
  ├── pr-test-analyzer
  └── custom agents...

Advanced Features Stack

┌─────────────────────────────────────────┐
│  CLI Commands (11 groups, 30+ commands) │
├─────────────────────────────────────────┤
│  Plugin System (5+ plugins)             │
│  - feature-development                  │
│  - pr-review-toolkit                    │
│  - commit-automation                    │
│  - security-hooks                       │
├─────────────────────────────────────────┤
│  Core Services                          │
│  - Multi-Agent Orchestration            │
│  - Security Scanner                     │
│  - Test Generator                       │
│  - Codebase Indexer                     │
│  - Web Search                           │
│  - Session Manager                      │
│  - Workflow Engine                      │
├─────────────────────────────────────────┤
│  AI Providers                           │
│  - OpenAI / Anthropic / Azure / Local   │
└─────────────────────────────────────────┘

Why Agentic CLI?

Comprehensive Feature Set

• 60+ commands covering every aspect of development
• 9 specialized AI agents for different tasks
• 5 production-ready plugins with extensible architecture
• Multiple AI provider support (not locked to one service)

Security-First Development

• Built-in vulnerability scanning (8+ types)
• Real-time security warnings during coding
• Auto-fix capabilities for common issues
• Compliance-ready security reports

Quality Automation

• Automated test generation with coverage analysis
• 6 specialized code review agents
• Error handling validation
• Code simplification suggestions

Advanced Automation

• Multi-step workflow orchestration
• Session management with audit trails
• Web search for grounded answers
• Codebase indexing for fast navigation

Development

# Development mode
npm run dev -- <command> [options]

# Watch for changes
npm run watch

# Run tests
npm test

# Lint and format
npm run lint
npm run format

# Validate before commit
bash scripts/validate.sh

Documentation

• README.md (this file) - Overview and getting started
• COMMAND_REFERENCE.md - Complete command documentation
• FINAL_SUMMARY.md - Project summary and statistics
• CONTRIBUTING.md - Contributing guidelines
• Plugin READMEs - Individual plugin documentation

Plugins Included

1. feature-development

7-phase guided development with code-architect, code-explorer, and code-reviewer agents.

2. pr-review-toolkit

6 specialized review agents for comprehensive code quality analysis.

3. commit-automation

Streamlined git workflows with AI-generated commit messages.

4. security-hooks

Real-time security validation with pre-edit and pre-commit hooks.

5. example-plugin

Template showing plugin development best practices.

License

MIT

Contributing

Contributions welcome! See CONTRIBUTING.md for guidelines.

Support

• GitHub Issues: https://github.com/judeotine/Agentic-CLI/issues
• Documentation: See docs in repository
• Command Help: agentic <command> --help

Technology Stack

Built with:

• TypeScript - Type-safe development
• Commander.js - CLI framework
• Zod - Runtime validation
• p-queue - Concurrency control
• simple-git - Git operations
• axios - HTTP client
• blessed - Terminal UI
• Multiple AI providers (OpenAI, Anthropic, Azure, Local)

Repository: https://github.com/judeotine/Agentic-CLI.git