The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses environment managers such as pixi, uv, or poetry to initialize projects, manage dependencies, and install the local package in editable mode.- [COMMAND_EXECUTION]: Generates and executes temporary Python scripts in a scratch/ directory for introspection tasks, such as extracting metrics from machine learning reports and updating project narratives.- [SAFE]: Implements comprehensive safety gates (G-TABULAR, G-PKG-NAME, G-ENV-MGR) and a mandatory pre-flight checklist to ensure user confirmation and project environment readiness before making changes.- [SAFE]: Utilizes predefined templates for scaffolding core project files and experiments, ensuring a standardized and predictable code structure.- [SAFE]: Strictly defines the responsibilities of each module (e.g., data.py, pipeline.py, evaluate.py) to prevent cross-contamination of logic and ensure experiment reproducibility.- [SAFE]: Processes local configuration files like pyproject.toml to detect existing project names and layouts, with mandatory user confirmation for any values used.

organize-ml-workspace