Home / Integrations / Plugins

🧩 Plugins

Code Puppy features a plugin system that allows extending functionality through callbacks. Plugins can intercept and modify behavior at various points in the execution flow.

Built-in Plugins

🔐 Claude Code OAuth

Authentication plugin for Claude API using OAuth flow. Enables secure authentication without storing API keys locally.

Location: plugins/claude_code_oauth/

💬 ChatGPT OAuth

Authentication plugin for ChatGPT/OpenAI using OAuth. Similar to Claude OAuth but for OpenAI services.

Location: plugins/chatgpt_oauth/

🛡️ Shell Safety

Safety plugin that evaluates shell commands for risk before execution. Blocks dangerous commands and requires confirmation for risky ones.

Location: plugins/shell_safety/

📁 File Permission Handler

Manages file permissions and access control. Ensures safe file operations with proper confirmations.

Location: plugins/file_permission_handler/

⚙️ Customizable Commands

Enables user-defined custom commands via markdown files. Create your own slash commands!

Location: plugins/customizable_commands/

Shell Safety Plugin

The shell safety plugin evaluates commands before execution:

🛡️ Risk Levels

LOW - Safe commands (ls, cat, echo)
MEDIUM - Moderate risk (git, npm, pip)
HIGH - Potentially dangerous (rm, sudo)
CRITICAL - Very dangerous (rm -rf, format)

Configure safety level:

/set safety_permission_level medium
# Options: low, medium, high, critical

Custom Commands Plugin

Create custom slash commands by adding markdown files to .claude/commands/:

# .claude/commands/deploy.md

## Deploy to Production

1. Run all tests
2. Build the production bundle
3. Deploy to the production server
4. Verify the deployment

Use the following commands:
```bash
npm test
npm run build
npm run deploy:prod
```

Then use with: /deploy

Creating Custom Plugins

Plugins use a callback system to hook into Code Puppy's execution flow:

# my_plugin/__init__.py

from code_puppy.callbacks import register_callback

@register_callback("before_shell_command")
async def my_command_hook(command: str, context: dict):
    """Called before every shell command."""
    # Modify or validate the command
    return {"allow": True, "command": command}

@register_callback("after_file_edit")
async def my_file_hook(file_path: str, content: str):
    """Called after every file edit."""
    # Log or post-process file changes
    print(f"File modified: {file_path}")

Available Callbacks

Callback	Description
`before_shell_command`	Before executing shell commands
`after_shell_command`	After shell command execution
`before_file_edit`	Before modifying files
`after_file_edit`	After file modifications
`before_message`	Before sending to AI
`after_response`	After receiving AI response
`on_startup`	When Code Puppy starts
`on_shutdown`	When Code Puppy exits

Plugin Structure

my_plugin/
├── __init__.py      # Plugin entry point with callbacks
├── config.py        # Plugin configuration
├── handlers.py      # Callback handlers
└── utils.py         # Utility functions