[Bug]: JSON schema conversion failed with Local OLLAMA or LM Studio

### Description

When using Local OLLAMA and LM Studio, other MCP servers such as Microsoft Learn are working fine but this Azure Resource Management MCP will throw the following error:
Execution failed: Error: Failed to get response from the AI model; retried 5 times (total retry wait time: 31.25 seconds) Last error: Engine protocol predict request returned 400: {"error":{"code":400,"message":"JSON schema conversion failed:\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true","type":"invalid_request_error"}}

### Steps to Reproduce

Github Copilot Desktop App
1. Add Azure Resource Management MCP - https://mcp.management.azure.com
2. Select Local Ollama or customer Endpoint to local Ollama or LM Studio
3. Ask the question which trigger using above MCP such as list out Azure SQL in subscription XXXX.

### Expected Behavior

List all Azure SQL in subscription XXXX

### Actual Behavior

Execution failed: Error: Failed to get response from the AI model; retried 5 times (total retry wait time: 31.25 seconds) Last error: Engine protocol predict request returned 400: {"error":{"code":400,"message":"JSON schema conversion failed:\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true","type":"invalid_request_error"}}

### VSCode Version

N/A

### ARM MCP Server Version

N/A

### Model Being Used

GPT-OSS, Gemma4:Latest, QWEN2

### Trace ID

_No response_

### Logs / Screenshots

```text
Execution failed: Error: Failed to get response from the AI model; retried 5 times (total retry wait time: 31.25 seconds) Last error: Engine protocol predict request returned 400: {"error":{"code":400,"message":"JSON schema conversion failed:\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true\nUnrecognized schema: true","type":"invalid_request_error"}}
```

### Additional Context

_No response_

Reproduce the failure through the Azure Resource Management MCP endpoint in GitHub Copilot Desktop using Local Ollama or LM Studio and one of the listed models. Trace where the MCP tool schemas become the model request, focusing on the repeated “Unrecognized schema: true” response; done means the Azure SQL request completes instead of failing after retries.

Written by the indexing model from the issue text.