Resources & Prompts

Resources Overview

Expose read-only data to your AI agents through the MCP resource system.

What are Resources?

Resources are read-only data sources exposed via the Model Context Protocol (MCP). They allow your AI agents to access structured data without modifying it. Resources can represent anything from database tables and files to external API responses and computed data.

Agent Dojo supports the following resource types:

database

Query results from a database table or view

file

Contents of a file or document

api

Response from an external API endpoint

static

Fixed content that rarely changes

dynamic

Content generated on each read request

function

Output of a serverless function invocation

Resource Properties

Every resource has the following properties:

Property	Type	Description
id	string	Unique identifier (UUID), auto-generated
server_id	string	ID of the server this resource belongs to
name	string	Human-readable name for the resource
uri	string	Unique URI identifying the resource (e.g., `db://users`)
description	string	Brief description of the resource and its contents
mime_type	string	MIME type of the resource content (e.g., `application/json`)
resource_type	string	One of: database, file, api, static, dynamic, function
requires_credentials	boolean	Whether the resource needs credentials to be read
cache_ttl_seconds	number	How long to cache the resource content (in seconds)
is_active	boolean	Whether the resource is currently active and available

Registering a Resource

curl -X POST https://api.agentdojo.dev/mcp/resources \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "server_id": "srv_001",
    "name": "User Directory",
    "uri": "db://users",
    "description": "Read-only access to the user directory",
    "mime_type": "application/json",
    "resource_type": "database",
    "requires_credentials": false,
    "cache_ttl_seconds": 300,
    "is_active": true
  }'

Listing Resources

List all resources registered on a server:

curl -X POST https://api.agentdojo.dev/mcp/resources/list \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "server_id": "srv_001"
  }'

Response:

{
  "resources": [
    {
      "uri": "db://users",
      "name": "User Directory",
      "description": "Read-only access to the user directory",
      "mimeType": "application/json"
    }
  ]
}

Reading a Resource

Read the contents of a specific resource by providing its URI:

curl -X POST https://api.agentdojo.dev/mcp/resources/read \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "uri": "db://users"
  }'

Response:

{
  "contents": [
    {
      "uri": "db://users",
      "mimeType": "application/json",
      "text": "[{\"id\": 1, \"name\": \"Alice\"}, {\"id\": 2, \"name\": \"Bob\"}]"
    }
  ]
}

Via MCP Protocol

Resources are also accessible through the standard MCP JSON-RPC protocol. AI agents typically use these methods to discover and read resources.

`resources/list`

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "resources/list"
}

`resources/read`

{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "resources/read",
  "params": {
    "uri": "db://users"
  }
}