Interactive PerturbNMF pipeline runner. Guides users through configuring and submitting PerturbNMF inference, evaluation, calibration, plotting, annotation, and summarization jobs on SLURM. Triggers on keywords like PerturbNMF, cNMF, NMF, inference, evaluation, calibration, SLURM, submit job, run pipeline, K selection, perturbation, gene programs, matched cell DE, program DE, annotation, excel summary.
Detect parameter drift across the PerturbNMF pipeline. Cross-checks every `--flag` defined in `src/**/*.py` (argparse) and `src/**/*.R` (optparse) against what is mentioned in sibling README.md files, sibling .sh runners, and the perturbNMF-runner skill markdown (SKILL.md + references/*.md). Use this skill whenever the user says "check drift", "drift check", "check parameter drift", "run drift check", "are the docs in sync", "did I update everything", "check if README/sh/skill match the .py", "sync check", or "doc sync". Also trigger after the user edits any pipeline argparse, README, SLURM .sh runner, or the perturbNMF-runner skill references — and any time they ask whether parameter names in .py, .sh, README, and the skill files agree.
Run the cNMF benchmarking pipeline test suite end-to-end: clean previous outputs, run sk-cNMF (CPU) and torch-cNMF (GPU via SLURM) inference tests, evaluation tests, and summarize pass/fail results. Use this skill whenever the user says "run tests", "run the tests", "test the pipeline", "rerun tests", "run sk-cNMF tests", "run torch tests", "run inference tests", "run evaluation tests", or any variation asking to execute, rerun, or check the cNMF test suite. Also trigger when the user asks to "clean and rerun", "test sk and torch", or "submit test jobs".
Inspect an .h5mu (MuData) file and generate a structured txt summary of its contents in a tree format. Lists all modalities, their X matrices, obs/var columns, uns, obsm, layers, obsp, varm, varp, and MuData-level attributes. Use this skill whenever the user asks to inspect, describe, summarize, or explore the structure of a .h5mu file, or when they want to know what's inside a MuData object saved on disk. Also trigger when the user says things like "what's in this h5mu", "show me the structure", "list the columns", "describe this mudata", or references .h5mu files and wants to understand their contents.