[BUG] Classification workflows are discoverable but not executable end-to-end

[SHIVANSH-ux-ys]

Reference Issues/PRs

Fixes #405

What does this implement/fix? Explain your changes.

Currently, agents can successfully discover classification estimators but fail when attempting to execute fit_predict because the executor blindly passes forecasting-specific arguments (like fh) to classifiers.

This PR updates executor.py to check the _estimator_type. If it is a "classifier", it safely routes the fitting and prediction logic to standard (X, y) parameters, omitting the incompatible fh argument.

Does your contribution introduce a new dependency? If yes, which one?

No new dependencies are introduced.

What should a reviewer concentrate their feedback on?

Reviewers can check the fit and predict methods in executor.py to ensure the classification routing logic covers all necessary edge cases.

Any other comments?

This is part of a larger ESoC stability push. I've separated this out into its own focused PR to make review as simple as possible.

PR checklist

• I have read the CONTRIBUTING guide.
• I have checked that my PR does not duplicate an existing PR.
• I have checked that there is an existing issue for this PR, if applicable.
• My code follows the project's coding standards.