mlflow
Installation
SKILL.md
MLflow
Official docs
| Topic | URL |
|---|---|
| Documentation | https://mlflow.org/docs/latest/index.html |
| Tracking | https://mlflow.org/docs/latest/tracking.html |
| Model Registry | https://mlflow.org/docs/latest/model-registry.html |
| Projects | https://mlflow.org/docs/latest/projects.html |
Agent rules
- Experiment naming — follow repo convention (
{project}/{env}/{task}). - Artifacts — log params, metrics, and artifacts; do not log secrets or raw PII.
- Registry stages — Staging → Production transitions need explicit promotion notes.
- Backend store — respect configured tracking URI (local, Databricks, remote server).
- Reproducibility — log git commit, conda/pip env, and data version when repo does.