SkyNet

A powerful local AI coding assistant CLI powered by Ollama. Like Claude Code, but running entirely on your own hardware.

Features

• Local LLM Inference - Runs on your GPU via Ollama (no API keys needed)
• Multi-Model Support - Switch between models on the fly (/models, /model)
• File Operations - Read, write, and edit files with intelligent search/replace
• Code Search - Grep and glob through your codebase
• Shell Execution - Run commands with safety guards
• Git Integration - Status, diff, commit, log, and branch management
• Web Search - Search the web and fetch documentation
• Session Persistence - Save and resume conversations
• Rich Terminal UI - Markdown rendering, syntax highlighting, and more
• Context Management - Auto-summarization to stay within context window limits
• Plan Mode - Create and approve execution plans before implementation
• Task Tracking - Track progress on multi-step tasks with the todo system

Requirements

• Python 3.11+
• Ollama installed and running
• A GPU with sufficient VRAM for your chosen model (e.g., RTX 3090/4090/5090 for 32B models)

Installation

# Clone the repository
git clone https://github.com/PySecNinja/skynet.git
cd skynet

# Install in editable mode
pip install -e .

# Or install directly
pip install .

Quick Start

# Make sure Ollama is running
ollama serve

# Pull a coding model (if you haven't already)
ollama pull qwen2.5-coder:32b

# Start SkyNet
skynet

Usage

Basic Commands

# Start interactive session
skynet

# Use a specific model
skynet --model qwen2.5-coder:7b

# Resume last session
skynet --resume
skynet -r

# Resume specific session
skynet --session 20231231_143022

# Skip confirmation prompts
skynet --no-confirm

In-Session Commands

Command	Description
/help	Show available commands
/models	List installed Ollama models
/model <name|number>	Switch to a different model
/sessions	List saved sessions
/resume <id>	Resume a specific session
/save	Manually save current session
/clear	Clear conversation history
/todos	Show current task list
/context	Show token usage and context window status
/plan	Enter/exit plan mode
/approve	Approve a pending execution plan
/reject	Reject a pending plan or exit plan mode
/quit, /exit, /q	Save and exit

Available Tools

SkyNet has access to 14 tools:

File Operations

• read_file - Read file contents with line numbers
• write_file - Create or overwrite files
• edit_file - Search and replace within files

Code Search

• grep - Search file contents with regex
• glob - Find files by pattern

Shell

• bash - Execute shell commands (with safety guards)

Git

• git_status - Show repository status
• git_diff - Show changes
• git_commit - Create commits
• git_log - View commit history
• git_branch - Manage branches

Web

• web_search - Search the web via DuckDuckGo
• web_fetch - Fetch and extract text from URLs

Planning & Tasks

• todo_write - Track progress on multi-step tasks
• create_plan - Create execution plans for approval in plan mode

Configuration

SkyNet can be configured via environment variables:

# Default model
export CLAUDE_CLONE_MODEL="qwen2.5-coder:32b"

# Ollama host
export CLAUDE_CLONE_OLLAMA_HOST="http://localhost:11434"

# Context window size
export CLAUDE_CLONE_NUM_CTX=32768

# Temperature
export CLAUDE_CLONE_TEMPERATURE=0.3

Recommended Models

Model	VRAM	Use Case
qwen2.5-coder:32b	~20GB	Best coding performance
qwen2.5-coder:14b	~10GB	Good balance
qwen2.5-coder:7b	~5GB	Fast, lightweight
deepseek-coder:33b	~20GB	Alternative coding model
llama3.1:8b	~5GB	General purpose

Project Structure

skynet/
├── pyproject.toml          # Project metadata and dependencies
├── README.md               # This file
├── LICENSE                 # MIT License
└── src/
    └── claude_clone/
        ├── cli.py          # CLI entry point and REPL
        ├── config.py       # Configuration management
        ├── core/
        │   ├── agent.py    # Main orchestration loop
        │   ├── context.py  # Context window management and auto-summarization
        │   ├── plan.py     # Plan mode state management
        │   ├── session.py  # Session persistence
        │   └── tokens.py   # Token counting utilities
        ├── llm/
        │   └── ollama_provider.py  # Ollama integration
        ├── tools/
        │   ├── base.py     # Tool base class
        │   ├── registry.py # Tool registration
        │   ├── file_ops.py # File tools
        │   ├── search.py   # Grep/glob tools
        │   ├── shell.py    # Bash tool
        │   ├── git.py      # Git tools
        │   ├── web.py      # Web search/fetch
        │   ├── plan.py     # Plan creation tool
        │   └── todo.py     # Task tracking tool
        └── ui/
            └── console.py  # Rich terminal UI

Plan Mode

Plan mode provides a structured workflow for complex tasks:

1. Enter plan mode with /plan
2. Explore the codebase - Only read-only tools are allowed (read, grep, glob, git status, etc.)
3. Create a plan - The AI uses create_plan to outline steps
4. Review and approve - Use /approve to proceed or /reject to discard
5. Execute - After approval, all tools are available for implementation

This prevents the AI from making changes before you've reviewed the approach.

> /plan
Plan mode activated. I'll explore and create a plan before making changes.

> Add user authentication to the app

# AI explores codebase, then creates a plan...

> /approve
Plan approved! Beginning execution...

Context Management

SkyNet automatically manages the context window to prevent token limit issues:

• Token Tracking - Use /context to see current usage
• Auto-Summarization - When context exceeds 75%, older messages are summarized
• Recent Preservation - Recent exchanges are kept intact for continuity

This allows for unlimited conversation length through automatic summarization.

Safety Features

• Dangerous Command Blocking - Prevents destructive commands like rm -rf /
• Confirmation Prompts - Asks before writing files or running risky commands
• Path Validation - Restricts access to sensitive directories
• Timeout Protection - Commands timeout after 2 minutes by default

Development

# Install with dev dependencies
pip install -e ".[dev]"

# Run linting
ruff check src/

# Run type checking
mypy src/

# Run tests
pytest

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Ollama for making local LLM inference easy
• Rich for beautiful terminal output
• Click for CLI framework
• Inspired by Claude Code by Anthropic

---

Note: This project is not affiliated with Anthropic or Claude. "SkyNet" is a playful name - this AI is here to help, not to take over the world!