Question: Should run_container_training validate container image references before submission?

## Description

The `run_container_training()` tool in `kubeflow_mcp/trainer/api/training.py` currently accepts the `image` argument and passes it directly to the Trainer SDK.

Empty or malformed image values may reach Kubernetes, where the failure appears later as an image-pull or container-start error. This makes the input problem difficult to identify before the job is submitted.

adding basic validation before the Trainer SDK call, such as:

- Rejecting empty or whitespace-only image values.
- Rejecting clearly malformed container image references.
- Returning a structured `VALIDATION_ERROR`.
- Ensuring the Kubernetes client is not called when validation fails.
Can i create issue and submit PR for this

Start in kubeflow_mcp/trainer/api/training.py at run_container_training() and inspect the existing image handoff and error handling. Define the validation scope from the listed requirements, then verify that invalid values return VALIDATION_ERROR and that the Kubernetes client is not called when validation fails.

Written by the indexing model from the issue text.