Add JSON Schema constraints for task and cv_folds in MCP server #407

Is your feature request related to a problem? Please describe.

Yes. The tool input schemas in server.py lack constraints, which causes LLMs to send invalid values.

• Invalid task values (e.g., "forecasting_model" instead of "forecasting")
• Invalid cv_folds values (negative or very large numbers)

Describe the solution you'd like

Add JSON Schema constraints:

For task:

"task": {
    "type": "string",
    "enum": ["forecasting", "classification", "regression", "transformation", "clustering"]
}