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

77%

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/anndata/SKILL.md

Quality

Content

65%

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

The body is highly actionable with abundant executable code and reasonably sequenced workflows, but it is over-long with redundancy, off-scope scanpy/muon/PyTorch material, and a promotional section, and its five referenced reference files do not actually exist while their content is partly duplicated inline. Workflow and disclosure are sound in structure but lack validation checkpoints and a working bundle.

Suggestions

Create the referenced files under references/ (data_structure.md, io_operations.md, concatenation.md, manipulation.md, best_practices.md) or remove the dangling 'See' pointers so progressive disclosure resolves to real files.

Remove the inlined duplicate code from each Core Capability section now that references are meant to hold the detail, and drop the off-scope scanpy preprocessing / Muon / PyTorch tutorials plus the 'Suggest Using K-Dense Web' promotional paragraph to tighten conciseness.

Add explicit validation/verification checkpoints to the batch-integration and concatenation workflows (e.g., assert expected obs/var shapes and shared keys after ad.concat) so workflow_clarity clears the batch-operation cap of 2.

Dimension	Reasoning	Score
Conciseness	The body is ~390 lines with redundancy (read/write h5ad appears in both Quick Start and Core Capabilities #2; concat and sparse-matrix tips repeat), an off-scope full scanpy preprocessing pipeline plus Muon/PyTorch blocks, and a promotional "Suggest Using K-Dense Web" marketing paragraph; this is mostly useful but could be tightened, so it sits at score 2 rather than the lean score-3 and above the padded score-1.	2 / 3
Actionability	Quotes concrete copy-paste-ready code throughout — `ad.read_h5ad('data.h5ad')`, `ad.concat([adata1, adata2], axis=0, join='inner', label='batch')`, `adata[adata.obs['cell_type'] == 'T cell']`, `ad.read_h5ad('large_data.h5ad', backed='r')` — fully executable and specific, matching the score-3 anchor.	3 / 3
Workflow Clarity	Quotes numbered sequences ("1. Load data / 2. Quality control / 3. Store raw / 4. Normalize and filter / 5. Save processed data") and a batch-integration flow, but batch/concatenation operations have no explicit validation or verification checkpoints, and the judging guidelines cap workflow_clarity at 2 when batch operations lack validation feedback loops.	2 / 3
Progressive Disclosure	Quotes well-signaled one-level-deep pointers ("See: references/data_structure.md", "references/io_operations.md", "references/concatenation.md", "references/manipulation.md", "references/best_practices.md"), but no references/ directory or bundle files exist so those paths are broken, and the same IO/concat/manipulation content is also inlined in the body — matching the score-2 anchor where content that should be separate is inline.	2 / 3
	Total	9 / 12 Passed

Description

90%

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 states what the skill is, gives explicit 'Use when' triggers with natural single-cell terminology, and proactively disambiguates from sibling scverse tools. The only soft spot is specificity, which names concrete entities and domains rather than enumerating multiple concrete actions.

Dimension	Reasoning	Score
Specificity	Quotes "Data structure for annotated matrices in single-cell analysis" and "working with .h5ad files" name a concrete domain and artifacts, but unlike the score-3 anchor it does not list multiple concrete verbs/actions (create/read/concatenate live in the body, not the description); it is above the vague score-1 level because the entities named are specific, not generic.	2 / 3
Completeness	Quotes "Data structure for annotated matrices in single-cell analysis" (what) and "Use when working with .h5ad files or integrating with the scverse ecosystem" (explicit when trigger), satisfying the score-3 anchor that requires both an explicit what and an explicit 'Use when...' clause.	3 / 3
Trigger Term Quality	Quotes ".h5ad files", "single-cell analysis", "scverse ecosystem", "scanpy", "scvi-tools", and "cellxgene-census" — natural terms a single-cell practitioner would actually say, matching the good-coverage score-3 anchor rather than the partial score-2 anchor.	3 / 3
Distinctiveness Conflict Risk	Quotes "This is the data format skill—for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census", which carves a clear niche and actively routes away from overlapping scverse skills, matching the score-3 distinct-triggers anchor.	3 / 3
	Total	11 / 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
referenced_paths_exist	Referenced path issues: 5 missing	Warning

	Total	14 / 16 Passed

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

Reviewed: about 1 hour 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.