| name | holomotion-train-interpreter |
| description | Run Python scripts or inspect installed packages using the HoloMotion training conda env. Use when running project Python scripts, executing training-related code, or inspecting packages in the holomotion_train environment. Always source train.env first, then use Train_CONDA_PREFIX. |
HoloMotion training interpreter
When to use
Invoke this skill when you need to:
- Run Python scripts with the project’s training environment (correct Python and deps)
- Execute training-related or holomotion Python code
- Inspect installed packages (e.g.
pip list, pip show, site-packages) in the training env
Workflow
- Source the environment so
Train_CONDA_PREFIX is set.
- Use that env’s Python/pip via
$Train_CONDA_PREFIX/bin/python and $Train_CONDA_PREFIX/bin/pip.
Where to source from:
- From repo root:
source holomotion/train.env
- From holomotion dir:
source train.env
Run a Python script (from holomotion dir):
source train.env && "$Train_CONDA_PREFIX/bin/python" path/to/script.py [args...]
Run a Python script (from repo root):
source holomotion/train.env && "$Train_CONDA_PREFIX/bin/python" holomotion/path/to/script.py [args...]
Inspect installed packages:
source train.env && "$Train_CONDA_PREFIX/bin/pip" list
source train.env && "$Train_CONDA_PREFIX/bin/pip" show <package>
One-off Python (e.g. check version or import):
source train.env && "$Train_CONDA_PREFIX/bin/python" -c "import sys; print(sys.executable)"
Rules
- Always source
train.env (or holomotion/train.env from repo root) before using Train_CONDA_PREFIX.
- Use
"$Train_CONDA_PREFIX/bin/python" and "$Train_CONDA_PREFIX/bin/pip" so the correct env is used; avoid relying on whatever python is on PATH.
- For package inspection, use the same env (e.g.
$Train_CONDA_PREFIX/bin/pip list or reading under $Train_CONDA_PREFIX/lib/python3.11/site-packages/).