cellxgene-census

Query the CELLxGENE Census (61M+ cells) programmatically. Use when you need expression data across tissues, diseases, or cell types from the largest curated single-cell atlas. Best for population-scale queries, reference atlas comparisons. For analyzing your own data use scanpy or scvi-tools.

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 ./skills/bio/cellxgene-census/SKILL.md

Quality

Content

71%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.

A highly actionable, code-rich skill body that clearly sequences query workflows with useful guardrails. Its main weaknesses are notable redundancy across the workflow, use-case, and best-practices sections, an off-topic promotional section, and broken/missing reference files that undermine progressive disclosure.

Suggestions

Either add the missing references/census_schema.md and references/common_patterns.md to the bundle, or remove the "Reference Documentation" section that promises them, so signaled navigation is not broken.

Consolidate the duplicated examples: fold "Common Use Cases" into "Core Workflow Patterns" (or move them into references/common_patterns.md) and remove best-practice items already stated inline to reduce redundancy.

Remove or relocate the K-Dense Web promotional section; it is unrelated to executing the skill and adds token cost without actionable value.

Dimension	Reasoning	Score
Conciseness	The code blocks are relevant and not concept-explaining, but "Common Use Cases" largely duplicates "Core Workflow Patterns", "Key Concepts and Best Practices" repeats inline points (is_primary_data appears 4+ times), and the K-Dense Web promo section is off-mission padding, so it could be tightened noticeably.	3 / 5
Actionability	Extensive copy-paste-ready, executable code covers the common cases (open_soma, get_anndata, get_obs/get_var, axis_query out-of-core, scanpy and PyTorch integration) with concrete filter syntax and parameters; the only placeholders are standard ML boilerplate Claude already knows.	5 / 5
Workflow Clarity	Numbered patterns (1-7) and an explicit "Two-Step Workflow: Explore Then Query" plus an "Estimate Query Size Before Loading" checkpoint give a clear sequence, and the Troubleshooting section provides error-recovery guidance, but validate/retry loops are not embedded as in-workflow checkpoints.	4 / 5
Progressive Disclosure	Sections are well headed and references are signaled with "When to read" notes, but the referenced files (references/census_schema.md, references/common_patterns.md) do not exist in the bundle and large reference-style material (metadata field tables, full code catalog) is inlined, so navigation is broken and content is not appropriately split.	3 / 5
	Total	15 / 20 Passed

Description

82%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, third-person description with explicit what/when triggers and helpful boundary guidance that distinguishes it from adjacent analysis skills. Its only weakness is that the action vocabulary is essentially a single verb (query) with scope elaboration rather than a list of distinct capabilities.

Suggestions

Add one or two more distinct concrete actions beyond "query" (e.g. "summarize cell-type composition", "download pre-calculated embeddings") to lift specificity from 3 toward 4-5.

Consider adding "scRNA-seq" as a synonym trigger term users commonly say alongside "single-cell".

Dimension	Reasoning	Score
Specificity	It names the domain and one core concrete action ("Query the CELLxGENE Census ... programmatically") elaborated by data scope (tissues, diseases, cell types), but does not list several distinct actions, so it sits at the 1-2-actions anchor rather than the multi-action anchor 4.	3 / 5
Completeness	It explicitly answers both what ("Query the CELLxGENE Census (61M+ cells) programmatically") and when ("Use when you need expression data across tissues, diseases, or cell types ..."), plus boundary guidance ("For analyzing your own data use scanpy or scvi-tools"), matching the explicit what-and-when anchor.	5 / 5
Trigger Term Quality	Natural user-facing terms are well covered ("expression data", "tissues", "diseases", "cell types", "single-cell atlas", "population-scale queries"), with only minor synonyms missing (e.g. "scRNA-seq"), matching the good-coverage anchor.	4 / 5
Distinctiveness Conflict Risk	It carves a clear niche (CELLxGENE Census, 61M+ cell single-cell atlas) with distinct triggers and explicit redirection to scanpy/scvi-tools for the user's own data, giving minimal conflict risk.	5 / 5
	Total	17 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (511 lines); consider splitting into references/ and linking	Warning
metadata_version	'metadata.version' is missing	Warning
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	13 / 16 Passed

Repository: wu-yc/LabClaw
Path: skills/bio/cellxgene-census/SKILL.md
Commit: df37802

Reviewed: 2 days 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.