Agent can generate identify anomalies for a given dataset

[mzbdraidia]

When asked to “detect anomalies” the system can:
- load dataset
- Use RAG to ground code generation
- Train anomaly detection model
- Return a table of anomalies

If a dataset is loaded it can use it to run the fit and detection, if not it will load a sample dataset