The Agent Skills Directory

[SAFE]: The skill serves as a legitimate technical guide for the Stable Baselines3 library, providing standard Python templates for reinforcement learning workflows.
[SAFE]: Dependency management follows best practices, referencing well-known and official packages from the Python ecosystem such as gymnasium, numpy, and stable-baselines3.
[SAFE]: Local file system operations (using os.makedirs and model saving/loading) are strictly confined to managing project-specific directories for logs, checkpoints, and evaluation results.
[SAFE]: The use of multiprocessing via SubprocVecEnv is a documented and standard approach for parallelizing reinforcement learning environments and does not represent an unauthorized privilege escalation or malicious persistence attempt.

stable-baselines3