Add Your Command

Contribute to the 1shot ecosystem by adding your own commands to the registry. Support standard prompts or enhanced MCP server integrations.

GitHub Repository

Fork and contribute to the registry

View Repository

Registry File

View the source registry directly

View registry.ts

MCP Server Support

NEW

1shot now supports MCP (Model Context Protocol) servers for enhanced functionality. MCP servers provide additional context and capabilities to your commands, enabling richer integrations with documentation, APIs, and specialized tools.

Currently Supported

assistant-ui

Full support for assistant-ui MCP documentation server, providing comprehensive integration guides and component documentation.

More MCP server integrations coming soon! We're actively working on expanding support.

Step-by-Step Guide

1

Edit the Registry

Open packages/1shot/src/registry.ts and add your command:

Basic Command Structure
export const registry: Record<string, RegistryEntry> = {
  // ... existing entries ...
  
  "your-command-name": {
    prompt: "Your main prompt that describes what the command does",
    systemPrompt: "Optional system prompt for additional context",
    // Optional: Add MCP server configuration
    mcpServers: {
      "server-name": {
        "command": "npx",
        "args": ["-y", "@your-org/mcp-server"]
      }
    }
  },
  
  // ... more entries ...
};
2

Command Examples

Here are different types of commands you can create:

Simple Command

"readme": {
  prompt: "Generate a comprehensive README.md file for this project"
}

Command with System Prompt

"stripe-integration": {
  systemPrompt: "You are an expert in Stripe payment integration",
  prompt: "Integrate Stripe payment processing into this application"
}

Command with MCP Server (Currently assistant-ui only)

"assistant-ui": {
  systemPrompt: "You are helping integrate assistant-ui into a project",
  prompt: "Integrate assistant-ui chat into this project",
  mcpServers: {
    "assistant-ui": {
      "command": "npx",
      "args": ["-y", "@assistant-ui/mcp-docs-server"]
    }
  }
}
3

Test Your Command

Build and test your command locally before submitting:

Build and test
# Install dependencies
npm install

# Build the project
npm run build

# Test your command
npx ./packages/1shot your-command-name
4

Submit a Pull Request

Once your command is working, commit your changes and create a pull request:

Submit your contribution
# Commit your changes
git add packages/1shot/src/registry.ts
git commit -m "Add [your-command-name] command to registry"

# Push to your fork
git push origin main

# Create a pull request on GitHub

PR Description Tips:

  • Describe what your command does
  • Include example use cases
  • Mention any dependencies or requirements
  • Note if you're using MCP servers

Best Practices

Clear Naming

Use descriptive, kebab-case names that clearly indicate the command's purpose (e.g., `stripe-checkout`, `jwt-auth`)

Comprehensive Prompts

Write detailed prompts that provide clear context and expectations for the AI assistant

Use System Prompts

Add system prompts for specialized knowledge or when the command requires specific expertise

Consider MCP Servers

For complex integrations, consider adding MCP server support (currently assistant-ui only) to provide enhanced documentation and context

Test Thoroughly

Always test your command in different project contexts before submitting

Ready to Contribute?

Join our community of developers building the future of AI-powered development tools.