The Agent Skills Directory

[SAFE]: The skill serves as documentation for MLflow, a widely recognized and trusted open-source tool for machine learning operations. All provided code samples and instructions follow standard, best-practice implementations of the MLflow framework.
[COMMAND_EXECUTION]: Includes standard commands for installing the tool via pip and running the local UI (mlflow ui). These are expected behaviors for a development-focused skill.
[EXTERNAL_DOWNLOADS]: Documents integration with trusted cloud services such as AWS SageMaker and Azure ML, and provides commands to push containers to official registries (e.g., ECR). These target well-known, trusted services.
[DATA_EXPOSURE]: Includes an example of starting a tracking server with a PostgreSQL backend (postgresql://user:password@localhost/mlflow). This is clearly marked as a placeholder for local configuration and is consistent with standard setup documentation.
[METADATA_POISONING]: All metadata fields (name, description, tags) accurately reflect the content and purpose of the skill without deceptive instructions.

mlflow