scanpy

Standard single-cell RNA-seq analysis pipeline. For quality control (QC), normalization, dimensionality reduction (PCA/UMAP/t-SNE), clustering, differential expression analysis, and visualization. Best suited for exploratory single-cell transcriptomics analysis using established workflows. For deep learning models, use scvi-tools; for data format issues, use anndata.

Quality

76%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./scientific-skills/Data Analysis/scanpy/SKILL.md

Quality

Content

60%Weight 40%Scale 1-5

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

The body delivers genuinely actionable scanpy guidance with a well-sequenced workflow and a properly documented bundle, but it is bloated: duplicated trigger sections, boilerplate padding, a broken internal reference, a dated path, and a promotional section unrelated to the skill. Trimming the inline tutorial in favor of the existing references would improve both conciseness and progressive disclosure.

Suggestions

Remove the duplicated 'When to Use' sections, the boilerplate 'Key Features'/'Implementation Details' blocks, and the K-Dense Web promotional section; keep one concise trigger list.

De-duplicate the inline step-by-step tutorial against references/standard_workflow.md — keep a short overview in SKILL.md and point to the reference for the full workflow.

Fix the incoherent navigation: the 'See ## Overview above' pointer targets a section that appears later, and the example path embeds a date ('20260316/...') that will rot.

Dimension	Reasoning	Score
Conciseness	Noticeably verbose with several padded sections: two overlapping 'When to Use' sections (one repeating the description in mangled lowercase), boilerplate 'Key Features'/'Implementation Details' filler, a dated example path ('cd "20260316/scientific-skills/..."'), a broken 'See ## Overview above' pointer, and an unrelated promotional K-Dense Web section. Much of the tutorial content (standard scanpy API usage) re-explains what Claude already knows.	2 / 5
Actionability	Concrete, executable code for every workflow stage and copy-paste-ready commands for the bundled script ('python scripts/qc_analysis.py input.h5ad --output filtered.h5ad --mt-threshold 5 --min-genes 200 --min-cells 3'). Minor gaps keep it below 5: the plotting section references an undefined 'genes' variable, and section 3 plots color='leiden' before Leiden clustering is introduced.	4 / 5
Workflow Clarity	The standard workflow is clearly sequenced (numbered steps 1-7) with most checkpoints present: QC violin plots before filtering, the PCA elbow plot before choosing n_pcs, trying multiple clustering resolutions, and 'save raw counts'/'save intermediate results' guidance. It falls short of 5 because there is no explicit validate-then-proceed feedback loop around the destructive filtering step.	4 / 5
Progressive Disclosure	All referenced bundle files exist (scripts/qc_analysis.py, references/standard_workflow.md, references/api_reference.md, references/plotting_guide.md, assets/analysis_template.py) and each is documented with its contents and a clear signal for when to read it, one level deep. Not 5: the inline step-by-step tutorial duplicates references/standard_workflow.md, and section organization is messy (duplicated When-to-Use sections, 'Implementation Details' referencing an Overview that appears later).	4 / 5
	Total	14 / 20 Passed

Description

83%Weight 40%Scale 1-5

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 description: it names a well-defined domain, enumerates concrete pipeline capabilities, states when it applies, and proactively disambiguates against neighboring tools (scvi-tools, anndata). The main gap is the absence of concrete 'Use when...' trigger phrases and file-format keywords (.h5ad, 10X) that would sharpen routing.

Suggestions

Add an explicit trigger clause, e.g., 'Use when analyzing single-cell RNA-seq data (.h5ad, 10X, CSV) or when the user mentions scRNA-seq, UMAP plots, cell clustering, or marker genes.'

Include common synonyms and file extensions (scRNA-seq, .h5ad, 10X Genomics) to raise trigger-term coverage for natural user phrasing.

Dimension	Reasoning	Score
Specificity	The description lists multiple concrete actions — 'quality control (QC), normalization, dimensionality reduction (PCA/UMAP/t-SNE), clustering, differential expression analysis, and visualization' — giving comprehensive coverage of the pipeline. It matches the 5 anchor (multiple specific concrete actions) and exceeds the 4 anchor, which is for coverage with minor gaps.	5 / 5
Completeness	Both what ('standard single-cell RNA-seq analysis pipeline' with enumerated steps) and when ('Best suited for exploratory single-cell transcriptomics analysis using established workflows') are present, plus boundary redirects to scvi-tools and anndata. The 'when' lacks concrete 'Use when...' trigger phrases, so it does not reach the 5 anchor; it is explicit rather than weakly implied, so above 3.	4 / 5
Trigger Term Quality	Strong natural keywords: 'single-cell RNA-seq', 'QC', 'clustering', 'UMAP', 't-SNE', 'differential expression'. Missing common variations and file formats users would say, such as 'scRNA-seq', '.h5ad', '10X', or 'marker genes', which keeps it below the 5 anchor's 'synonyms and file extensions' coverage.	4 / 5
Distinctiveness Conflict Risk	A clear single-cell RNA-seq niche with explicit conflict-avoidance guidance ('For deep learning models, use scvi-tools; for data format issues, use anndata') keeps overlap risk minimal, matching the 5 anchor's 'clear niche with distinct triggers'.	5 / 5
	Total	18 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: aipoch/medical-research-skills
Path: scientific-skills/Data Analysis/scanpy/SKILL.md
Commit: 686e09d

Reviewed: 13 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.