The Agent Skills Directory

[SAFE]: The skill consists of instructional content and Python templates for using the PyTorch Lightning framework. Analysis of all scripts and reference documentation shows no signs of malicious intent, data exfiltration, or prompt injection. The code provided is standard boilerplate for building neural networks, managing data modules, and configuring trainers.
[SAFE]: External dependencies referenced in the documentation, such as torch, lightning, wandb, and deepspeed, are well-known and standard libraries in the deep learning community. The skill does not automate their installation or execute remote scripts; they are mentioned only for documentation and configuration purposes.
[SAFE]: Sensitive information handling, such as API keys for experiment tracking services like WandB or Neptune, is documented using safe placeholders (e.g., YOUR_API_KEY), which aligns with standard development best practices.

pytorch-lightning