CodeUse Tool

A comprehensive code analysis and execution tool with modular architecture.

Project Structure & Team Ownership

tool/
├── main.py                    # 🎯 MAIN ENTRY POINT (Shared)
├── __init__.py               # 📦 PACKAGE INIT (Shared)
├── orchestrator/             # 👤 PERSON A - ORCHESTRATION
│   ├── loop.py               # Main workflow coordination
│   ├── state.py              # Application state management
│   └── prompts/
│       ├── __init__.py
│       └── system_prompts.py # LLM prompt templates
├── executor/                 # 👤 PERSON B - CODE EXECUTION
│   ├── __init__.py
│   └── executor.py           # Safe code execution engine
├── llm_parse/                # 👤 PERSON C - CODE ANALYSIS
│   ├── __init__.py
│   └── parser.py             # LLM-based code parsing
├── reporter/                 # 👤 PERSON D - REPORTING
│   ├── __init__.py
│   └── reporter.py           # Report generation
└── runs/                     # 📁 SHARED STORAGE
    └── .gitkeep             # Execution results & reports

Team Responsibilities

👤 Person A - Orchestrator

• Primary Focus: Workflow coordination and state management
• Key Files: orchestrator/loop.py, orchestrator/state.py
• Responsibilities:
  • Main application loop and workflow coordination
  • State management and data persistence
  • Integration between all modules
  • LLM prompt management

👤 Person B - Executor

• Primary Focus: Safe code execution and sandboxing
• Key Files: executor/executor.py
• Responsibilities:
  • Safe execution of user code
  • Multi-language support (Python, JavaScript, Bash, PowerShell)
  • Security and sandboxing
  • Resource monitoring and timeout handling

👤 Person C - LLM Parser

• Primary Focus: Code analysis and understanding
• Key Files: llm_parse/parser.py
• Responsibilities:
  • Code structure analysis
  • Function and class extraction
  • Issue detection and suggestions
  • Language-specific parsing logic

👤 Person D - Reporter

• Primary Focus: Report generation and output formatting
• Key Files: reporter/reporter.py
• Responsibilities:
  • Comprehensive report generation
  • Multiple output formats (JSON, Markdown)
  • Result visualization and formatting
  • Report management and retrieval

Getting Started

Environment Setup

1. Copy the environment template:

   cp env.example .env

2. Fill in your API keys in the .env file:

   # Required API Keys
   GEMINI_API_KEY=your_gemini_api_key_here
   OPENROUTER_API_KEY=your_openrouter_api_key_here
   OPENAI_API_KEY=your_openai_api_key_here
   
   # Optional Configuration
   OPENROUTER_MODEL=anthropic/claude-3.5-sonnet
   DEFAULT_TARGET_URL=http://localhost:5173
   EXECUTOR_BASE_URL=http://localhost:8001
   HTTP_REFERER=http://localhost

3. Install dependencies:

   pip install -r requirements.txt

Development

1. Each team member should focus on their assigned module
2. Use the existing interfaces and data structures
3. The main.py file coordinates all modules
4. Results are stored in the runs/ directory

Development Notes

• All modules are Python packages with proper __init__.py files
• Type hints and documentation are included for better collaboration
• Error handling and logging should be consistent across modules
• The orchestrator manages the flow between all components