The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to orchestrate the execution of machine learning training scripts (e.g., prepare_*_data.py and associated runners). It passes hyperparameters like learning rate and batch size as command-line arguments to these scripts.
[DATA_EXPOSURE]: The agent reads local training_history.json files generated by the training process. This is necessary for the skill to evaluate performance (MAE/loss) and decide on the next set of hyperparameters.
[SAFE]: No signs of prompt injection, data exfiltration, or persistence mechanisms were found. The skill operates on local datasets and history files as expected for an AutoML orchestrator.

ml-mlip-automl