[BUG] list_agents MCP tool violates OpenAI strict mode (object type without properties)

[pboers1988]

📋 Prerequisites

• I have searched the existing issues to avoid creating a duplicate
• By submitting this issue, you agree to follow our Code of Conduct
• I am using the latest version of the software
• I have tried to clear cache/cookies or used incognito mode (if ui-related)
• I can consistently reproduce this issue

🎯 Affected Service(s)

Controller Service

🚦 Impact/Severity

Minor inconvenience

🐛 Bug Description

When attempting to call list_agents from an MCP client like librechat using OpenAI models (strict mode on) on Foundry we receive the following error:

400 Invalid schema for function 'list_agents_mcp_<server>':                                                                                                                                                    
In context=(), object schema missing properties.

It seems like the mcp response is not setting the properties key inside the inputSchema object. Below a comparison between the list_agents tool and the invoke_agent tool.

🔄 Steps To Reproduce

1. Port-forward to the kagent-controller
2. Establish an MCP session
3. Observer the response of the MCP tools

List Agents

The output does not contain properties in inputSchema

  {                                                                                                                                                                                                                
    "name": "list_agents",                                                                                                                                                                                       
    "description": "List invokable kagent agents (accepted + deploymentReady)",
    "inputSchema": { "type": "object", "additionalProperties": false }
  }                                                                                                                                                                                                                

Compare to the well-formed invoke_agent entry on the same response, which has properties:

{                                                                                                                                                                                                              
  "name": "invoke_agent",
  "inputSchema": {
    "type": "object",
    "properties": { "agent": {...}, "task": {...}, "context_id": {...} },                                                                                                                                        
    "required": ["agent","task"],
    "additionalProperties": false                                                                                                                                                                                
  }                                                                                                                                                                                                            
}

🤔 Expected Behavior

a properties a object in the inputSchema, similar to the invoke_agent

📱 Actual Behavior

see reproduction

💻 Environment

Librechat: v8.6.0-rc0
Agentgateway: v1.2.1
Kagent: 0.9.4

❯ k version
Client Version: v1.36.1
Kustomize Version: v5.8.1
Server Version: v1.33.1

🔧 CLI Bug Report

No response

🔍 Additional Context

Architecture setup:

librechat -> agentgateway (auth handling) -> kagent-controller /mcp endpoint -> (6 agents)

📋 Logs

📷 Screenshots

No response

🙋 Are you willing to contribute?

• I am willing to submit a PR to fix this issue