home / mcp / context templates mcp server

Context Templates MCP Server

Provides access to reusable prompts and templates via an MCP endpoint and local npx execution.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "ginylil-context-templates": {
      "url": "https://detailer-api.supabase.co/functions/v1/mcp",
      "headers": {
        "AUTH_HEADER": "Bearer dtl_public_eec1f1037008dc82ce9d314c3294fbcc0e3f5d5df315d8c6"
      }
    }
  }
}

You can use the Context Templates MCP server to access reusable prompts, configurations, and workflows that streamline development tasks. It connects your MCP client to a remote service that processes templates and returns context-aware results to accelerate coding, planning, and automation tasks.

How to use

Connect your MCP client to the Context Templates server to access ready-to-use templates and execute them to generate feature plans, best practices, or task workflows. You can run templates remotely or pull specific templates for local customization, then execute them to receive immediate, context-aware outputs. Use the included tools to discover templates, retrieve raw content, or run templates directly to get generated results.

How to install

Prerequisites
- Node.js installed on your system
- An MCP client that supports HTTP or stdio MCP servers
- Access to the Context Templates MCP endpoint with a valid public token

Install and connect using the provided remote URL or the local npx-based command shown below.

{
  "mcpServers": {
    "context_templates_http": {
      "type": "http",
      "url": "https://detailer-api.supabase.co/functions/v1/mcp",
      "args": []
    }
  }
}

Context templates endpoints to use with MCP clients

Two primary ways to connect are provided: an HTTP-based remote endpoint and a local stdio-based command that runs via npx. Use the HTTP URL when your client supports remote MCP servers. Use the stdio configuration when you prefer to run the MCP locally via npx for quick experimentation or offline setups.

{
  "mcpServers": {
    "ctx_templates_http": {
      "type": "http",
      "url": "https://detailer-api.supabase.co/functions/v1/mcp",
      "args": []
    }
  }
}

Connecting with a local npx-based command

If you prefer a local, CLI-based approach, use the stdio configuration shown here. This runs the remote MCP via npx and passes the required authorization header.

{
  "mcpServers": {
    "ctx_templates_npx": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://detailer-api.supabase.co/functions/v1/mcp",
        "--header",
        "Authorization:${AUTH_HEADER}",
        "--allow-http"
      ],
      "env": {
        "AUTH_HEADER": "Bearer dtl_public_eec1f1037008dc82ce9d314c3294fbcc0e3f5d5df315d8c6"
      }
    }
  }
}

Configuration notes and practical tips

- After adding or editing MCP server configurations, restart your MCP client or IDE to load the new settings. The MCP server indicator should appear in your client after a successful connection.

- If you use the stdio approach, ensure Node.js is installed since the npx command relies on it. For the HTTP approach, no local runtime is required beyond your client’s support for MCP HTTP endpoints.

Security considerations

Keep your public token and authorization header secure. Do not expose the AUTH_HEADER in public repositories or shared configurations. Use environment variable substitution or secret management in your IDE or environment whenever possible.

Troubleshooting

If the MCP server indicator does not appear after configuration, verify that your client has network access to the endpoint, the JSON configuration is valid, and you have restarted the client entirely. Check that the AUTH_HEADER value is correctly provided if you are using the stdio configuration.

Available tools

list_templates

Discover all available templates, including their names, descriptions, and types.

get_template

Download a template’s raw content into your workspace for viewing or customization.

do_template

Execute a template directly to generate a context-aware output such as a feature plan or best practices.