scvi-tools

Probabilistic single-cell RNA-seq with scvi-tools — scVI for a batch-corrected latent space, scANVI for semi-supervised label transfer, and Bayesian differential expression. Reach for this skill to integrate scRNA-seq batches, embed cells for clustering, transfer annotations from a reference onto a query, or score differentially expressed genes per cluster. For spatial deconvolution / mapping use the cell2location, DestVI, or Tangram methods instead.

Quality

96%

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

Quality

Content

92%

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

A high-quality, actionable skill body with executable code and strong error-recovery guidance. The main weakness is progressive disclosure: no reference files exist and a referenced kernel.py is absent from the bundle.

Suggestions

Add the referenced kernel.py to a scripts/ (or skill) directory so the `exec(open("scvi-tools/kernel.py").read())` path resolves, or inline the h5ad_safe_obs helper directly in SKILL.md.

Move the full Modal pipeline.py example and the DE result-column listing into a references/ file, keeping SKILL.md as a lean overview with one-level-deep links.

Dimension	Reasoning	Score
Conciseness	Dense and operational — code blocks, gotchas, and troubleshooting tables with almost no explanatory padding; it scopes the library in one line ('wraps a family of deep generative models') and otherwise assumes Claude's competence.	3 / 3
Actionability	Fully executable, copy-paste-ready scVI/scANVI/DE code plus a concrete Modal pipeline.py with pinned versions ('scvi-tools==1.4.2', 'scanpy==1.11.5'); no pseudocode or vague direction.	3 / 3
Workflow Clarity	Clearly sequenced Setup→train→DE→output→remote sections, with gotchas and a symptom→fix troubleshooting table functioning as error-recovery feedback loops and an explicit ordering checkpoint ('preserve raw BEFORE any normalize/log1p').	3 / 3
Progressive Disclosure	Well-organized into clear sections, but everything is inline in one ~170-line file with no bundle files split out, and the referenced 'scvi-tools/kernel.py' path is not present in the provided bundle — so it is above the monolithic level 1 but below the well-split level 3.	2 / 3
	Total	11 / 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 tight, third-person description that names concrete capabilities, gives explicit positive and negative use-when triggers, and is well distinguished from adjacent skills. No verbosity or over-claims.

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions — 'scVI for a batch-corrected latent space, scANVI for semi-supervised label transfer, and Bayesian differential expression' — matching the top anchor rather than the single-action level 2.	3 / 3
Completeness	Explicitly answers both what (scVI/scANVI/Bayesian DE) and when ('Reach for this skill to integrate...'), and adds a negative trigger, so it is not capped at 2 for a missing use-when clause.	3 / 3
Trigger Term Quality	Uses natural analyst phrasings a user would actually say — 'integrate scRNA-seq batches, embed cells for clustering, transfer annotations from a reference onto a query, or score differentially expressed genes per cluster' — in third-person voice with good term coverage.	3 / 3
Distinctiveness Conflict Risk	Clear niche with distinct triggers, and the 'For spatial deconvolution / mapping use the cell2location, DestVI, or Tangram methods instead' clause actively routes adjacent tasks away, minimizing wrong-skill conflicts.	3 / 3
	Total	12 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: UnicomAI/wanwu
Path: configs/microservice/bff-service/configs/agent-skills/claude-science/scvi-tools/SKILL.md
Commit: f618458

Reviewed: about 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.