pipeline-drift-check SKILL.md skillA

pipeline-drift-check from engreitzlab/perturbnmf is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add engreitzlab/perturbnmf/pipeline-drift-check@git:20260929.add0212

Pins git:20260929.add0212 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
engreitzlab/perturbnmf
Path
.claude/skills/pipeline-drift-check/SKILL.md
Versions recorded
2
Latest version
git:20260929.add0212, committed Sep 29, 2026
First version recorded
Jun 9, 2026
Size
3,022 bytes
Repository stars
10
License
MIT

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…

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Pipeline Drift Check Skill

Detect param drift between source-of-truth argparse definitions and documentation/runner consumers in the PerturbNMF pipeline.

## What it checks

Source of truth: `add_argument("--flag", ...)` in every `src/**/*.py` and `make_option("--flag", ...)` in every `src/**/*.R`.

Consumers checked:
- **Sibling README.md** of each pipeline script — `--flag` mentions and `## Parameters` table rows must exist in that script's argparse.
- **Sibling .sh runner** — flags inside the `python3 <pipeline.py>` (or `Rscript`) block must exist in that script.
- **Skill markdown** (`.claude/skills/**/*.md`) — every `--flag` mentioned must be defined in *some* pipeline .py/.R, OR in a skill helper script under `.claude/skills/**/scripts/*.py`, OR in a test `conftest.py` `addoption()`.

## Suppressions (intentional, not bugs)

- Lines containing rename notes (`→`, `->`, "renamed", "formerly", "is now") — old names are documentation.
- Lines mentioning external CLI tools (`sacct`, `sbatch`, `jupyter`, `pytest`, etc.) — their flags belong to those tools.
- Wildcard prose placeholders like `--foo_*`.
- Skill convention placeholders: `--flag`, `--flag_name`, `---COMMENTED---`.
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add engreitzlab/perturbnmf/pipeline-drift-check@sha256:cfe4ff29de30a8af

The line at the top of this page pins the label git:20260929.add0212, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_qzsu6zvztxwowdbx.svg)](https://markdownregistry.com/a/art_qzsu6zvztxwowdbx)

3 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 2 versions of this file, between Jun 9, 2026 and Sep 29, 2026, about one every 112 days.

versioncommittedcommitsizeaudit
git:20260929.add0212 latest2026-09-29 add0212 3,022 BA view · diff
git:20260609.6e552382026-06-09 6e55238 3,210 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (3022 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

engreitzlab/perturbnmf · 10 stars · license MIT · pushed 2026-10-20 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_qzsu6zvztxwowdbx
GET https://markdownregistry.com/api/v1/resolve?ref=engreitzlab/perturbnmf/pipeline-drift-check
GET https://markdownregistry.com/api/v1/blob/cfe4ff29de30a8af8fd96e950aa8a41a9c30a8354ca68d58c4d11fbdb4b4df37

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from engreitzlab/perturbnmf

h5mu-structure SKILL.md skill
engreitzlab/perturbnmf · .claude/skills/h5mu-structure/SKILL.md · Inspect an .h5mu (MuData) file and generate a structured txt summary of its contents in a tree format. Lists all…
git:20260929.add0212 · audit A · 10 stars
perturbNMF-runner SKILL.md skill
engreitzlab/perturbnmf · .claude/skills/perturbNMF-runner/SKILL.md · Interactive PerturbNMF pipeline runner. Guides users through configuring and submitting PerturbNMF inference…
git:20260929.4c2f8fb · audit A · 10 stars
run-tests SKILL.md skill
engreitzlab/perturbnmf · .claude/skills/run-tests/SKILL.md · Run the cNMF benchmarking pipeline test suite end-to-end: clean previous outputs, run sk-cNMF (CPU) and torch-cNMF (GPU…
git:20260929.add0212 · audit A · 10 stars
CLAUDE.md CLAUDE.md file
engreitzlab/perturbnmf · CLAUDE.md
git:20260929.add0212 · audit A · 10 stars
DESIGN.md@src/Stage3_Interpretation/C_Annotation/4.Hypothesis/perturbseq-hypotheses DESIGN.md file
engreitzlab/perturbnmf · src/Stage3_Interpretation/C_Annotation/4.Hypothesis/perturbseq-hypotheses/DESIGN.md
git:20261019.294698d · audit A · 10 stars
perturbseq-hypotheses SKILL.md skill
engreitzlab/perturbnmf · src/Stage3_Interpretation/C_Annotation/4.Hypothesis/perturbseq-hypotheses/SKILL.md · Turn a finished Perturb-seq / PerturbNMF screen (cNMF programs, per-perturbation effect tables, NTC null, regulator…
git:20261019.294698d · audit A · 10 stars

Every file in engreitzlab/perturbnmf

Browse by kind, by grade A, or by owner.