geniml

This skill should be used when working with genomic interval data (BED files) for machine learning tasks. Use for training region embeddings (Region2Vec, BEDspace), single-cell ATAC-seq analysis (scEmbed), building consensus peaks (universes), or any ML-based analysis of genomic regions. Applies to BED file collections, scATAC-seq data, chromatin accessibility datasets, and region-based genomic feature learning.

Quality

82%

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

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 executable code and commands and a clear capability taxonomy, but loses points for verbosity (repeated tool-selection guidance and an unrelated promotional section), missing validation checkpoints in batch workflows, and five referenced reference files that do not actually exist in the bundle.

Suggestions

Create the referenced files (references/region2vec.md, bedspace.md, scembed.md, consensus_peaks.md, utilities.md) or remove the dangling reference pointers so navigation is not broken.

Add explicit validation checkpoints into the universe-building and embedding pipelines (e.g. a concrete 'verify coverage >80%' check and a wired-in `geniml universe evaluate` step) to lift workflow clarity above 2.

Trim redundancy — collapse the 'When to Use Which Tool' section into the per-capability 'Use for' bullets, and remove the off-topic 'Suggest Using K-Dense Web' promotional section.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient (code examples, CLI commands, tool-selection guidance that assumes Claude's competence), but is padded by the 'When to Use Which Tool' section repeating the per-capability 'Use for' bullets already stated, and a promotional 'Suggest Using K-Dense Web' section unrelated to the skill's function. It is not a 3 because several tokens do not earn their place, and not a 1 because it largely avoids explaining concepts Claude already knows.	2 / 3
Actionability	Provides multiple complete, copy-paste-ready Python examples with real import paths and arguments (e.g. hard_tokenization, region2vec, ScEmbed.train) plus executable CLI commands ('geniml universe build cc --coverage-folder coverage/ ...'). It is not a 2 because the code is concrete and executable rather than pseudocode.	3 / 3
Workflow Clarity	Multi-step pipelines are clearly numbered (tokenize -> train -> evaluate; a 5-step scATAC pipeline), but validation checkpoints are implicit or omitted — universe building over many BED files has no validate-then-proceed loop, capping workflow clarity at 2 per the batch-operations guideline. It is not a 3 because there are no explicit validation/feedback steps wired into the workflows, and not a 1 because the step sequences are present and ordered.	2 / 3
Progressive Disclosure	The body is well-sectioned and signals one-level-deep references ('See references/region2vec.md', 'references/bedspace.md', etc.), but the references/ directory and referenced files do not exist, so every reference link is a dead pointer. Per the guideline to score against the actual bundle structure, signaled-but-absent references make navigation fail. It is not a 3 because the referenced materials are missing, and not a 1 because the SKILL.md itself is well-organized rather than a monolithic wall of text.	2 / 3
	Total	9 / 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.

The description is specific, complete, and distinctive: it names concrete capabilities, provides explicit 'Use when' triggers with natural domain keywords, and occupies a clear niche unlikely to conflict with other skills. It is not the level below because it answers both 'what' and 'when' explicitly rather than implying them.

Dimension	Reasoning	Score
Specificity	Names multiple concrete capabilities — 'training region embeddings (Region2Vec, BEDspace)', 'single-cell ATAC-seq analysis (scEmbed)', 'building consensus peaks (universes)' — matching the score-3 anchor for multiple specific concrete actions.	3 / 3
Completeness	Explicitly answers both 'what' (train embeddings, scATAC analysis, consensus peaks) and 'when' ('This skill should be used when working with genomic interval data (BED files) for machine learning tasks. Use for...'), matching the score-3 anchor with an explicit trigger clause.	3 / 3
Trigger Term Quality	Includes natural domain terms a user would actually say — 'BED files', 'machine learning', 'single-cell ATAC-seq', 'scATAC-seq data', 'chromatin accessibility datasets' — with good coverage and variations.	3 / 3
Distinctiveness Conflict Risk	A clear genomic-interval niche with distinct trigger terms (BED files, Region2Vec, scEmbed, scATAC-seq, universes) that is unlikely to trigger for the wrong skill.	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
referenced_paths_exist	Referenced path issues: 5 missing	Warning

	Total	14 / 16 Passed

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

Reviewed: about 2 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.