Agentic CLI for software engineering automation. It features pluggable LLM providers, task interpretation, planning, execution, review and an agentic loop.
cargo install cli_engineer
cli_engineer --helpcli_engineer provides specialized commands for different types of software engineering tasks:
# Code generation
cli_engineer code "create a full-stack task management app with React frontend, Node.js/Express backend, PostgreSQL database, user authentication, and REST API"
# Code review and analysis
cli_engineer review "focus on error handling and performance"
# Documentation generation
cli_engineer docs
# Security analysis
cli_engineer security
# Refactoring assistance
cli_engineer refactor "improve code organization and performance"Dashboard UI (Default) - Interactive real-time interface with metrics and progress tracking:
# Dashboard with live metrics (default behavior)
cli_engineer docs "document the architecture"
# Dashboard + File Logging - logs appear in UI AND saved to file
cli_engineer -v security "analyze code security"
# Creates: cli_engineer_YYYYMMDD_HHMMSS.log with all session detailsSimple Text Mode - Traditional command-line output:
# Simple text output (disable dashboard)
cli_engineer --no-dashboard docs "document the architecture"
# Simple text + File logging
cli_engineer --no-dashboard -v review "analyze code quality"Key Features:
- 🎛️ Dashboard UI: Default experience with real-time metrics, progress bars, and live log display
- 📄 Simple Text: Clean terminal output for scripts or minimal environments
- 📝 File Logging: Verbose mode (
-v) automatically creates timestamped session logs - 🔄 Dual Output: Dashboard mode with verbose shows logs in UI AND saves to file simultaneously
For comprehensive guides and detailed information, see the docs/ directory:
- Documentation Home: Your starting point for all
cli_engineerdocumentation. - Quick Start: Get
cli_engineerup and running in minutes. - User Guide: Comprehensive guide to installation, configuration, and usage.
- Architecture Overview: Understand the system design, core components, and agentic loop.
- Command Reference: Detailed information on all CLI commands.
- LLM Providers: Information on supported Large Language Models.
- Contributing Guide: Learn how to contribute to the
cli_engineerproject.
| Command | Purpose | Output |
|---|---|---|
code |
Generate new code | Source files in current directory |
review |
Analyze existing code | code_review.md with findings |
docs |
Generate documentation | Documentation files in docs/ |
security |
Security analysis | security_report.md with vulnerabilities |
refactor |
Code improvement | Refactored source files |
git clone https://github.com/trilogy-group/cli_engineer
cd cli_engineer
cargo build
cargo test- 🤖 4 LLM Providers: OpenAI (Responses API), Anthropic (Claude 4), Google Gemini, Ollama (local)
- 🧠 Real-Time Thinking: Live reasoning traces from Claude 4, Gemini, and local models; reasoning summaries from o1/o3/o4-mini
- 📊 Dashboard UI (Default): Interactive interface with streaming thoughts and cost tracking
- 📝 Smart Buffering: Intelligent chunking at sentence boundaries for smooth reasoning display
- 💰 Accurate Costs: Real-time token usage and cost calculation from streaming events
- 🔄 Agentic Loop: Iterative planning, execution, and review with transparent reasoning
- 🔒 Local Option: Ollama support for privacy-focused, offline LLM inference (no API keys)
- 📁 Smart Artifacts: Context-aware file generation with proper restrictions
- 🔒 Command-Specific Behavior: Different file permissions and outputs per command type
- 📖 Comprehensive Documentation: Auto-generated docs with examples and API references
See the User Guide for detailed setup and advanced usage instructions.