The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill includes utility scripts (scripts/quick_train.py and scripts/visualize_filters.py) designed for local execution. These scripts interact with the file system to load datasets, save model checkpoints (.keras format), and output visualization images. This is standard behavior for machine learning workflows.
[DYNAMIC_EXECUTION] (INFO): The skill utilizes keras.models.load_model() and tf.function decorators. While model loading involves deserialization of model architectures and weights, it is a core requirement for the stated purpose of visualizing and resuming training of Keras models. Users should ensure that they only load model files from trusted sources.
[INDIRECT_PROMPT_INJECTION] (LOW): The scripts ingest user-provided model paths and training data. While this creates a theoretical surface for influence if an agent were to process malicious training logs or model metadata, the risk is negligible as the scripts perform standard numerical computations and do not incorporate untrusted natural language into system-level prompts.

deep-learning