Improve User Error Messages When MCP Description or Configuration is Incompatible with Kimi2.5

### Problem Description
When using Kimi2.5, you'll encounter errors if:
- The MCP configuration is too large, or any individual tool description is too long.
- The MCP tool schema/description format is not compatible with Kimi2.5 (for example, when the schema type or enum is invalid).

Example error messages:
```
[Error: Model call failed — BadRequestError:
Error code: 400 - {'error': {'message': "|nvalid request: tools.function.parameters is not a valid moonshot flavored json schema, details： ", 'type': 'invalid_request_error'}}
```

```
Error: Model call failed — BadRequestError: Error code: 400 - {'error': {'message': "Invalid request:
tools.function.parameters is not a valid moonshot flavored json schema, details: ", 'type':'invalid_request_error'}}
```
For end users, especially beginners, these messages are too technical and not user-friendly.

### Suggestions for Improvement
- Detect the root cause of incompatibility (e.g., size too large, invalid schema, incompatible field types) and show clear, actionable messages.
- Examples of user-friendly prompts:
- "The tool configuration is too large or contains unsupported types. Please reduce the number or length of tool descriptions and make sure all types are compatible with Kimi2.5."
- "Your tool schema contains an unsupported field type or an invalid enum. Please update your MCP tool configuration to match the requirements."
- Suppress raw schema errors, presenting only guidance the user can follow to correct their configuration.
- Optionally, validate MCP configuration/schema in advance and warn users of detected incompatibilities before invoking the model.

### Expected Outcome
- End users receive clear, friendly messages about configuration or schema errors, regardless of whether the cause is size, type, or schema incompatibility.
- Users are empowered to quickly fix their configuration without deciphering internal error details.

### Additional Information
This class of issues (size, format, schema compatibility) should all be grouped under helpful, actionable guidance for the user.