tooluniverse-single-cell

Single-cell RNA-seq analysis with scanpy/anndata — h5ad data loading, scRNA-seq quality control and QC gating (n_genes_by_counts, total_counts, mitochondrial percent / pct_counts_mt, pct_counts_ribo, doublet detection with Scrublet/scDblFinder, ambient RNA / SoupX awareness, empty-droplet filtering, MAD-based thresholds), normalization, dimensionality reduction (PCA, UMAP, t-SNE), clustering (Leiden, Louvain), marker gene identification, cell-type annotation, pseudotime/trajectory analysis. Use for any scRNA-seq workflow, including deciding which cells to filter, flag, or investigate before downstream analysis.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-sequenced skill with strong executable code and validation for the risky QC step. It is held back by some discursive sections and a broken reference to a missing `analysis_patterns.md` file.

Suggestions

Create `references/analysis_patterns.md` (or fix the references to point at an existing file) — it is cited for 'Pattern 1'/'Pattern 2' DE and correlation workflows but is absent from the bundle.

Tighten or relocate the discursive prose: collapse the QC biology explanations, 'Core Principles', 'Reasoning Framework', and 'Synthesis Questions' into the leanest actionable form or move them to a reference file.

Move the large ToolUniverse Integration tool list and the Scanpy-vs-Seurat table into reference files (the latter already overlaps `seurat_workflow.md`) to keep SKILL.md an overview.

Dimension	Reasoning	Score
Conciseness	The body is largely lean executable code and tables, but several sections are discursive and could be tightened — the QC biology prose ('A ruptured membrane lets cytoplasmic mRNA leak out...'), 'Core Principles', the 'Reasoning Framework for Result Interpretation', and the 'Synthesis Questions' list add explanatory bulk beyond what is strictly actionable.	2 / 3
Actionability	It provides abundant copy-paste-ready, executable code (data loading with orientation handling, QC metrics, the complete pipeline, statistical tests, Harmony batch correction) plus a concrete `tu run run_deseq2_analysis` command, fully meeting the executable-guidance anchor.	3 / 3
Workflow Clarity	A decision tree sequences the workflows, and the destructive QC-filtering operation has explicit validation/feedback: the install-plan preflight gate ('if scanpy is not installed... stop'), 'Always visualize distributions first', and per-step removal reporting — satisfying the validation-checkpoint requirement for risky operations.	3 / 3
Progressive Disclosure	References are one-level-deep and mostly real and well-signaled, but `analysis_patterns.md` is referenced three times (decision tree and Reference Documentation) yet does not exist in the bundle, and some inline content (the ToolUniverse Integration list, the Scanpy-vs-Seurat table duplicating `seurat_workflow.md`) could live in separate files.	2 / 3
	Total	10 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, third-person description that states concrete capabilities and an explicit 'Use for' trigger with natural keywords. It is specific, complete, and clearly distinguishable from adjacent skills.

Dimension	Reasoning	Score
Specificity	The description enumerates many concrete actions — 'h5ad data loading', 'scRNA-seq quality control and QC gating', 'normalization', 'dimensionality reduction (PCA, UMAP, t-SNE)', 'clustering (Leiden, Louvain)', 'marker gene identification', 'cell-type annotation', 'pseudotime/trajectory analysis' — matching the anchor for listing multiple specific concrete actions.	3 / 3
Completeness	It answers both 'what' (the analysis capabilities) and 'when' via the explicit trigger 'Use for any scRNA-seq workflow, including deciding which cells to filter, flag, or investigate before downstream analysis.'	3 / 3
Trigger Term Quality	It uses natural terms a user would say ('single-cell RNA-seq', 'scRNA-seq', 'h5ad', 'quality control', 'clustering', 'marker gene', 'trajectory') with good coverage and variations, not just technical jargon.	3 / 3
Distinctiveness Conflict Risk	It is tightly scoped to single-cell RNA-seq with scanpy/anndata, a clear niche with distinct triggers unlikely to fire for bulk RNA-seq or variant-analysis skills.	3 / 3
	Total	12 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: mims-harvard/ToolUniverse
Path: plugin/skills/tooluniverse-single-cell/SKILL.md
Commit: 089eb8e

Reviewed: about 9 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.