Skip to main content

​
Overview

The enkrypt_list_all_servers tool retrieves comprehensive information about all MCP servers configured in your gateway, including their tools, configuration status, and discovery requirements.
This is typically the first tool you’ll use when working with the gateway.

​
Tool Signature

async def enkrypt_list_all_servers(
    ctx: Context,
    discover_tools: bool = True
) -> dict

​
Parameters

​
ctx
Context
required
MCP context (automatically provided by the client)
​
discover_tools
bool
default:"true"
Whether to automatically discover tools for servers that need it.
  • true: Discovers and caches tools for all servers
  • false: Returns server list without tool discovery

​
Return Value

Returns a dictionary containing:
​
status
string
required
Operation status: "success" or "error"
​
available_servers
object
required
Dictionary of server configurations keyed by server name
​
servers_needing_discovery
array
List of server names that require tool discovery
​
discovery_failed_servers
array
List of server names where discovery failed (only when discover_tools=true)
​
discovery_success_servers
array
List of server names where discovery succeeded (only when discover_tools=true)

​
Usage Examples

​
Basic Usage

User: List all available servers

Claude: I'll list all available MCP servers for you.
[Uses enkrypt_list_all_servers tool]

I found 3 servers configured:
1. github - GitHub integration
2. weather - Weather data provider  
3. echo_server - Test echo server

​
Without Tool Discovery

If you want to list servers quickly without discovering their tools: 
List all servers without discovering tools
This will call enkrypt_list_all_servers with discover_tools=false, returning server information faster.

​
Response Structure

​
Success Response

{
  "status": "success",
  "message": "Tools discovery tried for all servers",
  "discovery_failed_servers": [],
  "discovery_success_servers": ["github", "weather", "echo_server"],
  "available_servers": {
    "github": {
      "server_name": "github",
      "description": "GitHub MCP Server",
      "config": {
        "command": "npx",
        "args": ["-y", "@modelcontextprotocol/server-github"]
      },
      "tools": {
        "create_issue": {
          "name": "create_issue",
          "description": "Create a new issue in a repository",
          "inputSchema": {
            "type": "object",
            "properties": {
              "repo": {"type": "string"},
              "title": {"type": "string"},
              "body": {"type": "string"}
            }
          }
        }
      },
      "input_guardrails_policy": {
        "enabled": true,
        "policy_name": "Sample Airline Guardrail"
      },
      "output_guardrails_policy": {
        "enabled": true,
        "policy_name": "Sample Airline Guardrail"
      }
    }
  }
}

​
Error Response

{
  "status": "error",
  "error": "No MCP config found. Please check your credentials.",
  "error_code": "CONFIG_MISSING_REQUIRED",
  "correlation_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890"
}

​
Server Object Structure

Each server in available_servers contains:
​
server_name
string
Unique identifier for the server
​
description
string
Human-readable description of the server
​
config
object
Server execution configuration
​
tools
object
Dictionary of available tools (empty if not discovered yet)
​
input_guardrails_policy
object
Input guardrail configuration for this server
​
output_guardrails_policy
object
Output guardrail configuration for this server

​
Implementation Details

​
Service Location

Implemented in: src/secure_mcp_gateway/services/server/server_listing_service.py:38

​
Authentication

The tool automatically:
  1. Extracts gateway credentials from the request context
  2. Validates credentials against the local or remote configuration
  3. Retrieves the MCP configuration for the authenticated user/project

​
Caching

Server information is cached to improve performance:
  • Cache Key: Based on gateway key, project ID, user ID, and config ID
  • Cache Duration: 24 hours (gateway config) + 4 hours (tools)
  • Cache Type: Local (in-memory) or External (Redis/KeyDB)

​
Parallel Processing

When discover_tools=true, the gateway:
  1. Phase 1: Validates all servers in parallel
  2. Phase 2: Separates servers with pre-configured tools from those needing discovery
  3. Phase 3: Runs tool validation and discovery in parallel
This parallel approach significantly reduces response time for multiple servers.

​
Tool Annotations

annotations={
    "title": "List Available Servers",
    "readOnlyHint": True,      # Does not modify state
    "destructiveHint": False,  # Not destructive
    "idempotentHint": True,    # Can be called multiple times
    "openWorldHint": False     # No external API calls
}

​
Common Use Cases

List all servers and show their configuration
Use this to verify that servers are properly configured and accessible.
What tools are available across all servers?
Gets a comprehensive view of all capabilities.
List servers without discovering tools
Quickly check server connectivity without waiting for tool discovery.
After setting up the gateway, use this tool to confirm all servers are accessible:
Show me all configured servers and whether they're working

​
Error Handling

Common errors and solutions:
Error: No MCP config foundSolution: Verify that:
  • Gateway key is correct in your MCP client config
  • ~/.enkrypt/enkrypt_mcp_config.json exists
  • Project ID and User ID match the configuration
Error: Authentication failedSolution: Check your gateway credentials in the MCP client configuration
Error: Tool discovery failed for some serversSolution:
  • Check server command and arguments
  • Verify required dependencies are installed
  • Review server logs for specific errors
  • Try discover_tools=false to see which servers are accessible

​
Performance Tips

Use Caching

Tools are cached for 4 hours by default. Subsequent calls return cached results instantly.

Skip Discovery

Set discover_tools=false for faster responses when you only need server names.

Parallel Execution

The gateway automatically discovers tools from multiple servers in parallel.

Monitor Metrics

Use enkrypt_get_timeout_metrics to identify slow servers.

Get Server Info

Get detailed information about a specific server

Discover Tools

Force re-discovery of tools for a server

Cache Status

View cache status for discovered tools

Execute Tools

Run tools from discovered servers

Build docs developers (and LLMs) love

Get started for freeTalk to us