anndata

Data structure for annotated matrices in single-cell analysis. Use when working with .h5ad files or integrating with the scverse ecosystem. This is the data format skill—for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census.

Quality

78%

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/anndata/SKILL.md

Quality

Content

68%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 well-structured, highly actionable skill body that uses references appropriately for depth. Its main weaknesses are verbosity from code duplicated across four sections and an off-topic promotional block, plus the absence of validation checkpoints in the batch and filtering workflows.

Suggestions

De-duplicate code: show read_h5ad / backed mode / sparse conversion once (e.g. in the IO and best-practices sections or their reference files) and remove the repeats from Quick Start, Common Workflows, and Troubleshooting.

Add explicit validation checkpoints to the workflows, e.g. after QC filtering print adata shape and after concatenation assert batch labels, so batch/destructive operations gain feedback loops.

Remove or move the 'Suggest Using K-Dense Web' promotional section out of the skill body; it is unrelated to AnnData usage and consumes context without earning its place.

Dimension	Reasoning	Score
Conciseness	Mostly efficient, executable content but padded by repeated code (read_h5ad / backed mode / sparse conversion appear in Quick Start, Core Capabilities, Common Workflows, and Troubleshooting) and a self-promotional 'Suggest Using K-Dense Web' section unrelated to the skill's function. Not a 4 because the duplication and promotional block are real, avoidable padding; not a 2 because the bulk is still concrete, useful guidance rather than concept explanation.	3 / 5
Actionability	Abundant copy-paste-ready, executable code covering creation, read/write, concat, subsetting, backed mode, chunked processing, and scanpy integration. Not below 5 because the common cases are covered with specific, runnable examples.	5 / 5
Workflow Clarity	Workflows are clearly numbered (e.g. Single-cell RNA-seq: Load → QC → Store raw → Normalize → Save) but lack explicit validation/verification checkpoints, and batch concatenation is a batch operation — the rubric caps such skills at 3. Not a 4 because no 'verify after filter' or 'confirm before writing' checkpoints are present; not a 2 because the sequences themselves are coherent and well-ordered.	3 / 5
Progressive Disclosure	Clean structure with well-signaled one-level-deep references — each Core Capability section lists 'See references/X.md' with a bullet summary, and all five referenced files exist (data_structure, io_operations, concatenation, manipulation, best_practices). Not a 5 because the body itself carries substantial inline code (Quick Start + per-section 'Common commands' + Common Workflows) beyond a lean overview, leaving minor redundancy with the reference files.	4 / 5
	Total	15 / 20 Passed

Description

87%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 that clearly states the data format's purpose, gives explicit 'Use when' trigger guidance, and proactively routes the user away from sibling scverse skills. Trigger coverage is good but could add 'RNA-seq' and 'zarr' for full synonym/extension completeness.

Dimension	Reasoning	Score
Specificity	Names the domain ('Data structure for annotated matrices in single-cell analysis') and concrete actions ('working with .h5ad files', 'integrating with the scverse ecosystem'), with only minor coverage gaps. Not a 5 because it states the format's purpose more than it enumerates multiple distinct concrete actions, and not a 3 because it goes beyond a single action with file-format and ecosystem triggers.	4 / 5
Completeness	Explicitly answers both 'what' ('Data structure for annotated matrices in single-cell analysis') and 'when' ('Use when working with .h5ad files or integrating with the scverse ecosystem') with concrete trigger phrases. Not below 5 because both halves are clearly and explicitly present.	5 / 5
Trigger Term Quality	Includes natural triggers users would say — 'single-cell analysis', '.h5ad files', 'scverse ecosystem', plus tool names 'scanpy', 'scvi-tools', 'cellxgene-census' — with good coverage and one file extension. Not a 5 because synonyms like 'RNA-seq'/'scRNA-seq' and extensions like 'zarr' are absent.	4 / 5
Distinctiveness Conflict Risk	Has a clear niche ('This is the data format skill') and explicitly disambiguates from related skills ('for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census'), giving minimal conflict risk. Not below 5 because the routing guidance makes overlap unlikely.	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
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: googolme/run0204
Path: scientific-skills/anndata/SKILL.md
Commit: 4902409

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.